The impact of new tetralogy drugs on ventricular remodeling in heart failure patients: a systematic review and network meta-analysis Authors: Xi-Wen Wang, Hao Li, Gan-Qi Wang, Ya-Nan Sheng, Xia Wang, Hui Pan, Zhi-Hao Wang, Wei Zhang, Yuan-Yuan Shang, Ming Zhong Published: 2025 ABSTRACT AbstractBackgroundHeart failure, particularly heart failure with reduced ejection fraction (HFrEF), represents a major global health challenge due to its high prevalence, rapid progression, and substantial mortality rates. While the “New Tetralogy” drugs (angiotensin receptor-neprilysin inhibitors, sodium-glucose cotransporter 2 inhibitors, beta-blockers, and mineralocorticoid receptor antagonists) have significantly advanced HFrEF management, their comparative effects on ventricular remodeling remain unclear. This network meta-analysis systematically evaluates these therapeutic agents to inform optimal clinical decision-making.MethodsWe conducted a comprehensive search of PubMed, Cochrane Library, Embase, and Web of Science for randomized controlled trials (RCTs) evaluating ventricular remodeling parameters in HFrEF patients. Twenty-two RCTs involving 16,425 participants were included. Data were analyzed using Stata 14.2 and RevMan 5.3, with outcomes expressed as mean differences (MD) and 95% confidence intervals (CI).ResultsOur study showed markedly different therapeutic characteristics: sodium-glucose cotransporter 2 inhibitors significantly reduced left ventricular mass index (MD = -6.57, 95% CI: -11.98 to -1.16), while beta-blockers demonstrated superior improvement in left ventricular ejection fraction compared to mineralocorticoid receptor antagonists (MD = 3.54, 95% CI: 0.60 to 6.48) through indirect comparison. Mineralocorticoid receptor antagonists showed greater efficacy in reducing left ventricular end-systolic volume (MD = -22.10, 95% CI: -31.42 to -12.78). Angiotensin receptor-neprilysin inhibitors outperformed renin-angiotensin system inhibitors across multiple parameters, including left ventricular end-systolic volume (MD = -7.00, 95% CI: -13.91 to -0.09), end-systolic volume index (MD = -13.78, 95% CI: -25.98 to -1.58), and end-diastolic volume index (MD = -5.80, 95% CI: -9.76 to -1.84).ConclusionsThis network meta-analysis demonstrates that each component of the “New Tetralogy” applies unique beneficial effects on ventricular remodeling. Angiotensin receptor-neprilysin inhibitors, sodium-glucose cotransporter 2 inhibitors, and mineralocorticoid receptor antagonists predominantly improve ventricular volumes, while beta-blockers significantly enhance systolic function. These findings provide evidence-based guidance for personalized therapeutic strategies in HFrEF management, particularly for patients with persistent left ventricular systolic dysfunction.Clinical trial numberNot applicable.Registration numberCRD42023454737.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12872-025-05173-x.Keywords: Heart failure with reduced ejection fraction, New tetralogy drugs, Ventricular remodeling, Randomized controlled trials, Network meta-analysis Introduction Heart failure has emerged as a global health crisis, characterized by high prevalence, rapid progression, and substantial mortality rates. Epidemiological data reveal a worldwide prevalence of 1–3% among adults. In China, 1.3% of adults aged 35 + are affected, with a 4.1% in-hospital mortality rate [1, 2]. Notably, heart failure with reduced ejection fraction (HFrEF) constitutes 40.2% of all heart failure cases, presenting with more severe clinical manifestations, impaired cardiac function, reduced exercise tolerance, and diminished quality of life compared to other heart failure subtypes. As the terminal stage of various cardiovascular diseases, heart failure remains a formidable challenge in cardiovascular medicine, necessitating comprehensive strategies for secondary prevention, cardiovascular rehabilitation, and quality of life improvement to ensure optimal long-term patient care. Ventricular remodeling, a pathophysiological adaptation to myocardial injury or increased cardiac load, manifests through structural alterations in ventricular size, morphology, wall thickness, and tissue composition, ultimately leading to ventricular dilation or hypertrophy. This complex process involves pathological cardiomyocyte hypertrophy, apoptosis, myofibroblast proliferation, and interstitial fibrosis, mediated by cellular interactions among cardiomyocytes, fibroblasts, endothelial cells, and leukocytes in response to cardiac injury, neurohormonal activation, and stress responses [3, 4]. Though initially compensatory, ventricular remodeling progressively causes left ventricular failure [2]. Consequently, interventions targeting ventricular remodeling are crucial for delaying disease progression, reducing cardiovascular complications, and improving patient outcomes. The integration of advanced imaging modalities, particularly echocardiography and cardiac magnetic resonance imaging, has revolutionized clinical practice by enabling precise evaluation of ventricular structure and function, facilitating accurate disease assessment and treatment planning. Pharmacological management remains the cornerstone of heart failure treatment. The conventional therapeutic triad (beta-blockers [BBs], renin-angiotensin system inhibitors [RASis], and mineralocorticoid receptor antagonists [MRAs]) has been expanded with the inclusion of sodium-glucose cotransporter-2 inhibitors (SGLT2is), forming the “New Tetralogy” of heart failure therapy. In recent years, multiple clinical studies [5–8] have confirmed that SGLT2is not only effectively reduce rehospitalization rates in patients with HFrEF, but also significantly improve patient outcomes and reduce all-cause mortality. The 2021 European Society of Cardiology (ESC) guidelines indicate that SGLT2is demonstrate significant clinical benefits in the treatment of patients with HFrEF. Growing evidence demonstrates that the “New Tetralogy” medications significantly improve cardiovascular outcomes by modulating ventricular remodeling processes [9–13]. For instance, carvedilol [14], a third-generation BB, has shown superior efficacy in reducing left ventricular volume and improving ejection fraction compared to placebo, while also attenuating left ventricular dilation. However, current research primarily focuses on placebo-controlled studies, leaving a critical gap in direct comparative effectiveness data among ARNIs, BBs, MRAs, and SGLT2is. This network meta-analysis aims to systematically evaluate the efficacy of these four therapeutic classes in managing ventricular remodeling in chronic heart failure patients, thereby providing evidence-based guidance for clinical decision-making and therapeutic optimization. Methods Search strategy Our search strategy followed the PRISMA extension guidelines for network meta-analysis [15, 16]. A systematic search was conducted on PubMed, Cochrane Library, Embase, and Web of Science to identify RCTs examining the impact of “New Tetralogy” therapy on ventricular remodeling in patients with HFrEF. The search was limited to English language and included literature published from the inception of the databases up to February 2025. The articles were searched using a combination of subject keywords and free text, incorporating references that were retrospectively included in the literature. The search language for this study was restricted to English. The following keywords and MeSH terms were applied: Heart Failure, Cardiac Failure, Heart Decompensation, Decompensation, Heart, Myocardial Failure, Left Sided Heart Failure, Right Sided Heart Failure, Ventricular Remodeling, Cardiac Remodeling, Ventricular, Ventricular Cardiac Remodeling, Myocardial Remodeling, Ventricular, Ventricular Myocardial Remodeling, Left Ventricle Remodeling, Left Ventricular Remodeling, Angiotensin-Converting Enzyme Inhibitors, Inhibitors, Angiotensin Converting Enzyme, Inhibitors, Kininase II, Kininase II Antagonists, Kininase II Inhibitors, Angiotensin Converting Enzyme Antagonists, ACE Inhibitors, Inhibitor, ACE, Angiotensin I Converting Enzyme Inhibitor, Antagonists, Angiotensin Converting Enzyme, Angiotensin II Type 1 Receptor Blockers, Angiotensin II Type 1 Receptor Blockers, Type 1 Angiotensin Receptor Blockers, Angiotensin II Type 1 Receptor Antagonist, Selective Angiotensin II Receptor Antagonists , Sartans, sacubitril and valsartan sodium hydrate drug combination, sacubitril valsartan sodium hydrate, sacubitril and valsartan drug combination, sacubitril-valsartan, sacubitril valsartan sodium anhydrous, sacubitril-valsartan sodium anhydrous drug combination, LCZ 696, Entresto, Adrenergic beta Antagonists, Adrenergic beta-Receptor Blockader, Blockader, Adrenergic beta-Receptor, beta Adrenergic Antagonist, beta Adrenergic Blocker, beta Adrenoceptor Antagonists, Adrenergic beta Receptor Blockaders, beta Adrenergic Receptor Blockaders, beta Adrenergic Blocking Agents, beta Adrenergic Receptor Blockader, Adrenergic beta Antagonist, Sodium-Glucose Transporter 2 Inhibitors, SGLT 2 Inhibitors, Sodium Glucose Transporter 2 Inhibitor, Inhibitor, SGLT2, Mineralocorticoid Receptor Antagonists, Mineralocorticoid Antagonists, Aldosterone Receptor Antagonists, Aldosterone Antagonists. Eligibility criteria Trials meeting the following criteria were included in this analysis: (1) study population: inclusion criteria required LVEF < 40%, while no particular New York Heart Association (NYHA) functional classification was mandated for enrollment; (2) intervention group and control group: patients in the intervention group were treated with a single agent from the “New Tetralogy” drugs, whereas controls received either another agent from this regimen or placebo alone; (3) reported N-terminal B-type natriuretic peptide (NT-proBNP) levels or at least one of the pre-specified indicators of left cardiac structure and function, assessed either by cardiac magnetic resonance imaging or echocardiography, including left ventricular mass index (LVMI), left ventricular end-systolic volume (LVESV), left ventricular end-diastolic volume (LVEDV), left ventricular end-systolic volume index (LVESVI), left ventricular end-diastolic volume index (LVEDVI), and LVEF; (4) study design: either primary or secondary analyses of RCTs. Literature was independently screened, extracted, and cross-checked by two researchers. In cases of disagreement, both parties negotiated and reached agreement or sought assistance from a third party. Duplicates were removed using EndNote X9 software. Titles and abstracts were initially screened to eliminate irrelevant literature, followed by a thorough reading of the full text to exclude literature that did not meet the criteria for inclusion. Data extraction Two reviewers independently extracted data from the included studies. Disagreements were resolved through consultation with a third reviewer to reach a consensus. The key information extracted from the study included baseline patient information such as the first author, year of publication, country, and demographic characteristics of participants, as well as details on subgroups, interventions, and outcome metrics. Study endpoints The primary outcomes assessed included LVMI, LVESV, LVEDV, LVESVI, and LVEDVI. The secondary outcomes comprised LVEF and NT-proBNP. Assessment of risk of bias The quality of the literature was assessed using the Cochrane Risk of Bias Assessment Tool [15], which evaluated various aspects such as the correct generation of randomization allocation sequence, concealment of randomization sequence, blinding of subjects, investigators, and assessors, completeness of outcome data, selective reporting of endpoints, and other potential sources of bias. Each aspect was categorized as low risk, high risk, or unclear risk of bias. Risk of bias plots were generated using RevMan 5.3 software to present the results. Data synthesis and analysis The data analysis and graphing for this study’s literature were performed using RevMan 5.3 and Stata 14.2, adhering to the frequency-based network meta-analysis framework. Continuous variables served as outcome indicators, with MD and 95% CI chosen for effect size measurement. Heterogeneity among studies was evaluated using Stata 14.2. A P value > 0.1 and I² ≤ 50% indicated no statistical heterogeneity, prompting the use of a fixed-effect model. Conversely, a P value ≤ 0.1 and I² >50% signaled heterogeneity, necessitating a random-effect model. For significant heterogeneity (I² >75%), further exploration of heterogeneity sources and subgroup analysis were recommended. Evidence maps visually depict the relationships among interventions utilizing Stata 14.2. Node sizes reflect sample sizes, while line thicknesses indicate the strength of direct evidence. Closed loops in these diagrams undergo inconsistency tests. If direct and indirect comparisons yield nonsignificant differences (P > 0.05), a consistency model is applied; otherwise, an inconsistency model with node-splitting is used to pinpoint local inconsistencies (P < 0.05). When no closed loops exist, the consistency model directly analyzes outcome indicators. The SUCRA score ranks these indicators. For studies exceeding 10, Egger’s test assesses small sample effects and publication bias; fewer than 10 studies preclude such an assessment. Results A total of 6,452 related documents were retrieved, including 1,183 from PubMed, 3,845 from Embase, 1,039 from Web of Science, and 385 from the Cochrane Library. Ultimately, 22 studies [6, 14, 17–36] were included in the reticulation meta-analysis, all of which were RCTs. The flowchart depicting the literature screening process can be found in Fig. 1. Flow chart of literature screening Characteristics of the included studies and risk of bias assessment Among the 22 RCTs analyzed, there were a total of 16,425 patient cases included. The publication period of these studies ranged from 1992 to 2022, with interventions involving RASi/ARNI, BB, MRA, SGLT2i, and placebo. Detailed characteristics of the included literature can be found in Tables 1 and 2, while the quality assessment of the studies is presented in Supplementary Fig. 1. Table 1Main characteristics of patients with HFrEF in the selected RCTs with available informationStudiesCountryTotal SampleMean Age (years)SexMale/FemaleMean BMI (kg/m²)Follow-up (weeks)Omar-2021 [21]Denmark19064.0(11)162/2829.512Shah-2022 [32](PARADISE-MISub study)America45764(12)74/26NA32Lee-2021 [6]Scotland10568.7(11.1)77/2830.736Jensen-2022 [19] (Empire-HF)Denmark12068(10)99/2129.512Santos-Gallego-2021 [31]America8462(12.1)54/3029.724Jensen-2020 [36](Empire-HF)Denmark19064.1(24.8)162/2829.612Tsutamoto −2004 [33]Japan3763(3.3)28/9NA16Edes-2005 [18]German26072.08(5.10)191/6926.0732Robert N-1997 [14]New ZealandAustralia12368.4(6.6)107/16NA48Palazzuoli-2002 [28]Italy4871(5)30/18NA48Pouleur-1993 [29]Belgium6556(9.5)54/11NA48Konstam-1992 [20]America11557.3104/11NA52Udelson-2010 [34]America22662.7(12.5)189/37NA36Leonetti L-2004 [22]Italy4076.89(5.9)28/12NA12Bellenger-2004 [17]England3471(8.5)NANA24Barr-1995 [23]England4269(3)32/10NA8Vizzardi-2010 [35]Italy15860(13)132/26NA24Rosenberg-2008 [30]Denmark40562(3)244/161NA24Packer-2020 [27](EMPEROR-Reduced trial)International373066.8(11)2837/89327.952Butt-2021 [24]England474266.3(10.9)3633/110928.232McMurray-2019 [25](DAPA-HF trial)England474466.3(10.9)3635/110928.173Nassif-2019 [26] (DEFINE-HF Trial)America51061.3(11.5)193/7031.512 Main characteristics of patients with HFrEF in the selected RCTs with available information Shah-2022 [32] (PARADISE-MI Sub study) Jensen-2020 [36] (Empire-HF) New Zealand Australia Packer-2020 [27] (EMPEROR-Reduced trial) McMurray-2019 [25] (DAPA-HF trial) Table 2Basic characteristics of the selected RCTs with available informationStudiesInterventionComparatorOther Therapies (%)History of Cardiovascular diseaseOutcomes VariablesConditionOmar-2021 [21]Empagliflozin(SGLT2i)PlaceboACEI/ARB(65%) ARNI(65%)MRA(66%)BB(95%)Atrial fibrillation(35%)Hypertension(40%)Ischemic heart disease(54%)LVESV LVEDV LVESVIHFrEFShah-2022 [32](PARADISE-MI Sub study)Sacubitril/Valsartan(ARNI)Ramipril(ACEI)ACEI/ARB (100%)MRA(57.5%)BB(100%)Myocardial infarction (22.3%)PCI(19.9%)CABG(5%)Hypertension(77.5%)Diabetes(51.6%)LVMILVEF LVESV LVEDV LVESVI LVEDVIHFrEFLee-2021 [6]Empagliflozin(SGLT2i)PlaceboARNI(34.3%)ACEI (46.7%)BB(91.4%)ARB(14.3%) MRA(60%)ACEI/ARB/ARNI(95.2%)Hypertension(70.5%)Coronary artery disease(70.5%)Prior myocardial infarction(83.8%)LVMI LVESV LVEDV LVESVILVEFNT-proBNPT2DM or prediabetesHFrEFJensen-2022 [19](Empire-HF)Empagliflozin(SGLT2i)PlaceboARNI(37.5%)BB(96.7%) MRA(69.2%)ACEI/ARB/ARNI(95%)Atrial fibrillationor flutter(59.8%)Ischemic heartdisease(40.8%)ICD (13.3%) CRT(25.9%)LVEFHFrEFTsutamoto −2004 [33]Spironolactone(MRA)PlaceboACEI(73.0%)BB(37.8%)Dilated cardiomyopathy (75.7%)Hypertensive heart disease (16.2%)Valvular heart disease (8%)LVEDV LVEFHFrEFBarr-1995 [23]Spironolactone(MRA)PlaceboNACoronary artery disease(100.0%)Prior myocardial infarction(60.0%)LVEFHFrEFVizzardi-2010 [35]Spironolactone(MRA)PlaceboACEI (88.0%)BB(89.2%)ARB(9.0%)Diabetes(20.3%)Hypertension(26.0%)Coronary artery disease(44.3%)Ischemic heart disease(55.7%)LVMI LVEF LVESV LVEDVHFrEF Basic characteristics of the selected RCTs with available information Empagliflozin (SGLT2i) ACEI/ARB(65%) ARNI(65%) MRA(66%) BB(95%) Atrial fibrillation(35%) Hypertension(40%) Ischemic heart disease(54%) Shah-2022 [32] (PARADISE-MI Sub study) Ramipril (ACEI) ACEI/ARB (100%) MRA(57.5%) BB(100%) Myocardial infarction (22.3%) PCI(19.9%) CABG(5%) Hypertension (77.5%) Diabetes(51.6%) LVMI LVEF LVESV LVEDV LVESVI LVEDVI Empagliflozin (SGLT2i) ARNI(34.3%) ACEI (46.7%) BB(91.4%) ARB(14.3%) MRA(60%) ACEI/ARB/ARNI (95.2%) Hypertension (70.5%) Coronary artery disease(70.5%) Prior myocardial infarction(83.8%) LVMI LVESV LVEDV LVESVI LVEF NT-proBNP T2DM or prediabetes HFrEF Jensen-2022 [19] (Empire-HF) Empagliflozin (SGLT2i) ARNI(37.5%) BB(96.7%) MRA(69.2%) ACEI/ARB/ARNI (95%) Atrial fibrillation or flutter(59.8%) Ischemic heart disease(40.8%) ICD (13.3%) CRT(25.9%) Spironolactone (MRA) ACEI(73.0%) BB(37.8%) Dilated cardiomyopathy (75.7%) Hypertensive heart disease (16.2%) Valvular heart disease (8%) Spironolactone (MRA) Coronary artery disease(100.0%) Prior myocardial infarction(60.0%) Spironolactone (MRA) ACEI (88.0%) BB(89.2%) ARB(9.0%) Diabetes(20.3%) Hypertension (26.0%) Coronary artery disease(44.3%) Ischemic heart disease(55.7%) Intercomparison and consistency testing of interventions Network maps of the meta-analysis and mixed comparison forest plots Supplementary material S1 presents a plot illustrating the mesh relationships among various interventions before and after the decomposition of RASi/ARNI. In this plot, the vertices represent different interventions, with their sizes corresponding to the sample sizes. A connecting line between two vertices signifies a direct comparison between the respective interventions, while the thickness of the line indicates the number of studies that have compared the two interventions. It was observed that the LVESVI network relationship graphs exhibited closed loops both before and after the splitting of RASi/ARNI. The Network package was utilized to test for inconsistencies, yielding a P value of 0.5594 for LVESVI before and after the splitting of RASi/ARNI. Subsequently, the node splitting method was employed to assess local inconsistency, revealing that all P values were greater than 0.05. This finding indicates the absence of local inconsistency, suggesting improved consistency and supporting the comparability of the groups. Network relationship evidence plots for the remaining outcome indicators did not form a closed loop; therefore, the consistency model was uniformly applied to analyze all outcome indicators. The results of the forest plot for the mixed comparisons (both direct and indirect comparisons) across the various interventions are presented in Fig. 2. Forest plots for (A) left ventricular mass index (LVMI), (B) left ventricular end-systolic volume (LVESV), (C) left ventricular end-diastolic volume (LVEDV), (D) left ventricular end-systolic volume index (LVESVI), (E) left ventricular end-diastolic volume index (LVEDVI), and (F) left ventricular ejection fraction (LVEF) RASi: renin-angiotensin system inhibitor; ACEI: angiotensin-converting enzyme inhibitor; ARB: angiotensin-receptor blocker; ARNi: angiotensin receptor–neprilysin inhibitor; BB: beta-blocker; MRA: mineralocorticoid receptor antagonist; SGLT2: sodium‒glucose cotransporter protein 2 inhibitor Results of network meta-analysis for each outcome indicator Network meta-analysis of LVMI Descriptive statistics of LVMI were analyzed in five included papers, comprising four interventions and a total of 838 patients. The findings of a meta-analysis on LVMI across different interventions revealed that SGLT2i led to a reduction in LVMI compared to placebo (MD = −6.57, 95% CI:−11.98, −1.16). Additionally, indirect comparisons suggested that the differences in LVMI among BB versus SGLT2i, SGLT2i versus MRA, and BB versus MRA were not statistically significant (Fig. 2A; Table 3). Table 3League table representing relative effect estimates for LVMIBB−5.42(24.27, 13.43)SGLT2i−9.99(−29.86, 9.88)−4.57(−14.47, 5.33)MRA−11.99(30.05, 6.06)−6.57(−11.98, −1.16)−2.00(−10.29, 6.29)Placebo League table representing relative effect estimates for LVMI −5.42 (24.27, 13.43) −9.99 (−29.86, 9.88) −4.57 (−14.47, 5.33) −11.99 (30.05, 6.06) −6.57 (−11.98, −1.16) −2.00 (−10.29, 6.29) Network meta-analysis of LVESV LVESV was statistically described in seven included studies, with five interventions and 1157 patients. The results of the network meta-analysis of LVESV under various interventions indicate that MRA significantly reduced LVESV (MD = −30.10, 95% CI: −36.59, −23.61), while BB also contributed to a reduction in LVESV (MD = −8.00, 95% CI: −14.69, −1.31) compared to placebo (Fig. 2B; Table 4). The SUCRA ordination plot further illustrates the effectiveness of these interventions, ranking LVESV in descending order as follows: MRA (SUCRA = 99.9%), BB (SUCRA = 52.9%), RASi/ARNI (SUCRA = 49.4%), SGLT2i (SUCRA = 40.4%), and placebo (SUCRA = 7.4%) (Supplementary Fig. 3B). Table 4League table representing relative effect estimates for LVESVMRA−22.10(−31.42, −12.78)BB−22.10(−37.45, −6.75)0.00(−15.44, 15.44)RASi/ARNI−24.18(−37.00, −11.37)−2.08(−15.00, 10.83)−2.08(−19.85, 15.68)SGLT2i−30.10(−36.59, −23.61)−8.00(−14.69, −1.31)−8.00(−21.92, 5.91)−5.92(−16.96, 5.13)Placebo League table representing relative effect estimates for LVESV −22.10 (−31.42, −12.78) −22.10 (−37.45, −6.75) 0.00 (−15.44, 15.44) −24.18 (−37.00, −11.37) −2.08 (−15.00, 10.83) −2.08 (−19.85, 15.68) −30.10 (−36.59, −23.61) −8.00 (−14.69, −1.31) −8.00 (−21.92, 5.91) −5.92 (−16.96, 5.13) Indirect comparative analyses of interventions after decomposing RASi/ARNI followed by MRA, demonstrated a significant reduction in LVESV compared to BB, RASi, and SGLT2i [(MD = −22.10, 95% CI: −31.42, −12.78); (MD = −22.10, 95% CI: −37.45, −6.75); (MD = −24.18, 95% CI: −37.00, −11.37)]. Furthermore, ARNI was found to be more effective in reducing LVESV compared to RASi (MD = −7.00, 95% CI: −13.91, −0.09) (Fig. 2B; Table 5). The SUCRA results, listed in descending order, were as follows: MRA (SUCRA = 99.2%), ARNI (SUCRA = 72.1%), BB (SUCRA = 46.1%), RASi (SUCRA = 40.1%), SGLT2i (SUCRA = 36.2%), and placebo (SUCRA = 6.3%) (Supplementary Fig. 4A). Table 5League table representing relative effect estimates for LVESV after decomposing RASi/ARNIMRA−15.10(−31.93, 1.73)ARNI−22.10(−31.42, −12.78)−7.00(−23.91, 9.91)BB−22.10(−37.45, −6.75)−7.00(−13.91, −0.09)−0.00(−15.44, 15.44)RASi−24.18(−37.00, −11.37)−9.08(−28.14, 9.98)−2.08(−15.00, 10.83)−2.08(−19.85, 15.68)SGLT2i−30.10(−36.59, −23.61)−15.00(−30.53, 0.53)−8.00(−14.69, −1.31)−8.00(−21.91, 5.91)−5.92(−16.96, 5.13)Placebo League table representing relative effect estimates for LVESV after decomposing RASi/ARNI −15.10 (−31.93, 1.73) −22.10 (−31.42, −12.78) −7.00 (−23.91, 9.91) −22.10 (−37.45, −6.75) −7.00 (−13.91, −0.09) −0.00 (−15.44, 15.44) −24.18 (−37.00, −11.37) −9.08 (−28.14, 9.98) −2.08 (−15.00, 10.83) −2.08 (−19.85, 15.68) −30.10 (−36.59, −23.61) −15.00 (−30.53, 0.53) −8.00 (−14.69, −1.31) −8.00 (−21.91, 5.91) −5.92 (−16.96, 5.13) Network meta-analysis of LVEDV LVEDV was statistically described in 8 included studies, comprising 5 interventions and 1194 patients. A network meta-analysis of LVEDV across various interventions revealed that the differences in outcomes among the different interventions were not statistically significant (Fig. 2C; Table 6). Additionally, further analysis following the decomposition of RASi/ARNI confirmed that the differences in outcomes among the various interventions remained statistically insignificant (Fig. 2C; Table 7). Table 6League table representing relative effect estimates for LVEDVSGLT2i0.16(−41.52, 41.84)RASi/ARNI−1.68(−33.70, 30.35)−1.84(−43.40, 39.73)MRA−3.85(−42.89, 35.20)−4.01(−51.19, 43.18)−2.17(−41.08, 36.74)BB−10.84(−33.58, 11.90)−11.00(−45.92, 23.92)−9.16(−31.70, 13.37)−6.99(−38.73, 24.74)Placebo League table representing relative effect estimates for LVEDV 0.16 (−41.52, 41.84) −1.68 (−33.70, 30.35) −1.84 (−43.40, 39.73) −3.85 (−42.89, 35.20) −4.01 (−51.19, 43.18) −2.17 (−41.08, 36.74) −10.84 (−33.58, 11.90) −11.00 (−45.92, 23.92) −9.16 (−31.70, 13.37) −6.99 (−38.73, 24.74) Table 7League table representing relative effect estimates for LVEDV after decomposing RASi/ARNIARNI−10.13(−62.41, 42.14)SGLT2i−10.00(−41.64, 21.64)0.13(−41.48, 41.75)RASi−11.81(−63.98, 40.36)−1.67(−33.69, 30.34)−1.81(−43.29, 39.67)MRA−13.97(−70.75, 42.80)−3.84(−42.90, 35.23)−3.97(−51.11, 43.17)−2.16(−41.11, 36.78)BB−20.97(−68.04, 26.09)−10.84(−33.58, 11.90)−10.97(−45.82, 23.87)−9.17(−31.70, 13.36)−7.00(−38.77, 24.76)Placebo League table representing relative effect estimates for LVEDV after decomposing RASi/ARNI −10.13 (−62.41, 42.14) −10.00 (−41.64, 21.64) 0.13 (−41.48, 41.75) −11.81 (−63.98, 40.36) −1.67 (−33.69, 30.34) −1.81 (−43.29, 39.67) −13.97 (−70.75, 42.80) −3.84 (−42.90, 35.23) −3.97 (−51.11, 43.17) −2.16 (−41.11, 36.78) −20.97 (−68.04, 26.09) −10.84 (−33.58, 11.90) −10.97 (−45.82, 23.87) −9.17 (−31.70, 13.36) −7.00 (−38.77, 24.76) Network meta-analysis of LVESVI LVESVI was statistically described in nine included studies, comprising five interventions and 1689 patients. Meta-analysis of LVESVI across different interventions indicated that BB reduced LVESVI compared to placebo (MD = −11.16, 95% CI: −22.02, −0.29). Additionally, indirect comparisons revealed no statistically significant differences in LVESVI between BB RASi/ARNI, SGLT2i, or MRA (Fig. 2D; Table 8). Table 8League table representing relative effect estimates for LVESVIBB−1.88(−5.72, 1.97)RASi/ARNI−9.46(−22.24, 3.31)−7.59(−20.34, 5.16)SGLT2i−11.16(−22.02, −0.29)−9.28(−20.59, 2.03)−1.69(−8.20, 4.82)Placebo−13.46(−30.73, 3.82)−11.58(−29.14, 5.98)−3.99(−18.92, 10.94)−2.30(−15.73, 11.13)MRA League table representing relative effect estimates for LVESVI −1.88 (−5.72, 1.97) −9.46 (−22.24, 3.31) −7.59 (−20.34, 5.16) −11.16 (−22.02, −0.29) −9.28 (−20.59, 2.03) −1.69 (−8.20, 4.82) −13.46 (−30.73, 3.82) −11.58 (−29.14, 5.98) −3.99 (−18.92, 10.94) −2.30 (−15.73, 11.13) Further analysis of RASi/ARNI decomposition revealed that ARNI significantly reduced LVESVI compared to the placebo group (MD = −13.78, 95% CI: −25.98, −1.58). The indirect results corroborated these findings, indicating that the differences in outcomes among the six interventions were not statistically significant when compared (Fig. 2D; Table 9). The SUCRA results, presented in descending order, were as follows: ARNI (SUCRA = 93.7%), BB (SUCRA = 77.4%), RASi (SUCRA = 58.5%), SGLT2i (SUCRA = 32.7%), and placebo (SUCRA = 20.2%) (Supplementary Fig. 4C). Table 9League table representing relative effect estimates for LVESVI after decomposing RASi/ARNIARNI−2.62(−8.60, 3.35)BB−4.50(−9.07, 0.07)−1.88(−5.72, 1.97)RASi−12.09(−25.64, 1.46)−9.46(−22.24, 3.31)−7.59(−20.34, 5.16)SGLT2i−13.78(−25.98, −1.58)−11.16(−22.02, −0.29)−9.28(−20.59, 2.03)−1.69(−8.20, 4.82)Placebo−16.08(−34.22, 2.06)−13.46(−30.73, 3.82)−11.58(−29.14, 5.98)−3.99(−18.92, 10.94)−2.30(−15.73, 11.13)MRA League table representing relative effect estimates for LVESVI after decomposing RASi/ARNI −2.62 (−8.60, 3.35) −4.50 (−9.07, 0.07) −1.88 (−5.72, 1.97) −12.09 (−25.64, 1.46) −9.46 (−22.24, 3.31) −7.59 (−20.34, 5.16) −13.78 (−25.98, −1.58) −11.16 (−22.02, −0.29) −9.28 (−20.59, 2.03) −1.69 (−8.20, 4.82) −16.08 (−34.22, 2.06) −13.46 (−30.73, 3.82) −11.58 (−29.14, 5.98) −3.99 (−18.92, 10.94) −2.30 (−15.73, 11.13) Network meta-analysis of LVEDVI LVEDVI was statistically described in 7 studies involving 5 interventions and 1109 patients. The results of a meta-analysis comparing LVEDVI under different interventions did not show any statistically significant differences (Fig. 2E; Table 10). Table 10League table representing relative effect estimates for LVEDVIBB0.84(−28.91, 30.60)RASi/ARNI−8.30(−24.28, 7.67)−9.15(−42.26, 23.97)SGLT2i−8.73(−21.25, 3.80)−9.57(−37.32, 18.19)−0.42(−9.03, 8.19)Placebo−12.03(−31.81, 7.76)−12.87(−44.57, 18.83)−3.72(−21.29, 13.84)−3.30(−18.61, 12.01)MRA League table representing relative effect estimates for LVEDVI 0.84 (−28.91, 30.60) −8.30 (−24.28, 7.67) −9.15 (−42.26, 23.97) −8.73 (−21.25, 3.80) −9.57 (−37.32, 18.19) −0.42 (−9.03, 8.19) −12.03 (−31.81, 7.76) −12.87 (−44.57, 18.83) −3.72 (−21.29, 13.84) −3.30 (−18.61, 12.01) Further analysis following the decomposition of RASi/ARNI revealed that ARNI significantly reduced LVEDVI compared to RASi (MD = − 5.80, 95% CI: −9.76, −1.84). However, the difference was not statistically significant when comparing the outcomes of the other interventions (Fig. 2E; Table 11). Table 11League table representing relative effect estimates for LVEDVI after decomposing RASi/ARNIARNI−6.64(−36.66, 23.37)BB−5.80(−9.76, −1.84)0.84(−28.91, 30.59)RASi−14.95(−48.29, 18.40)−8.30(−24.28, 7.67)−9.15(−42.25, 23.96)SGLT2i−15.37(−43.40, 12.66)−8.73(−21.25, 3.80)−9.57(−37.32, 18.18)−0.42(−9.03, 8.19)Placebo−18.67(−50.61, 13.27)−12.03(−31.81, 7.76)−12.87(−44.56, 18.82)−3.72(−21.28, 13.84)−3.30(−18.61, 12.01)MRA League table representing relative effect estimates for LVEDVI after decomposing RASi/ARNI −6.64 (−36.66, 23.37) −5.80 (−9.76, −1.84) 0.84 (−28.91, 30.59) −14.95 (−48.29, 18.40) −8.30 (−24.28, 7.67) −9.15 (−42.25, 23.96) −15.37 (−43.40, 12.66) −8.73 (−21.25, 3.80) −9.57 (−37.32, 18.18) −0.42 (−9.03, 8.19) −18.67 (−50.61, 13.27) −12.03 (−31.81, 7.76) −12.87 (−44.56, 18.82) −3.72 (−21.28, 13.84) −3.30 (−18.61, 12.01) Network meta-analysis of LVEF The analysis of LVEF in the 15 papers reviewed included 5 interventions and 1791 patients. Meta-analysis results indicated that BB and SGLT2i significantly increased LVEF compared to placebo (MD = 3.75, 95% CI: 1.61, 5.89; MD = 2.46, 95% CI: 0.00, 4.91). Indirect comparisons indicated that BB significantly elevated LVEF compared to MRA(MD = 3.54, 95% CI: 0.60, 6.48). However, no statistically significant differences were observed between the other comparisons (Fig. 2F; Table 12). When ranking the interventions based on the surface under the cumulative ranking curve, the results were as follows, in descending order: BB (SUCRA = 88.4%), SGLT2i (SUCRA = 67.2%), RASi/ARNI (SUCRA = 56.5%), MRA (SUCRA = 22.5%), and placebo (SUCRA = 15.3%) (Supplementary Fig. 3F). The results of further analysis and comparison following the disaggregation of RASi/ARNI were consistent with the findings prior to disaggregation (Fig. 2F; Table 13). The SUCRA ordinanation diagrams ranked the treatments in descending order: BB (SUCRA = 82.6%), ARNI (SUCRA = 69.1%), SGLT2i (SUCRA = 61.8%), RASi (SUCRA = 50.9%), MRA (SUCRA = 20.7%), and placebo (SUCRA = 14.9%) (Supplementary Fig. 4E). Table 12League table representing relative effect estimates for LVEFBB1.29(−1.96, 4.55)SGLT2i1.71(−2.79, 6.22)0.42(−4.25, 5.09)RASi/ARNI3.54(0.60, 6.48)2.24(−0.93, 5.42)1.82(−2.63, 6.28)MRA3.75(1.61, 5.89)2.46(0.00, 4.91)2.03(−1.94, 6.01)0.21(−1.81, 2.23)Placebo League table representing relative effect estimates for LVEF 1.29 (−1.96, 4.55) 1.71 (−2.79, 6.22) 0.42 (−4.25, 5.09) 3.54 (0.60, 6.48) 2.24 (−0.93, 5.42) 1.82 (−2.63, 6.28) 3.75 (1.61, 5.89) 2.46 (0.00, 4.91) 2.03 (−1.94, 6.01) 0.21 (−1.81, 2.23) Table 13League table representing relative effect estimates for LVEF after decomposing RASi/ARNIBB0.51(−5.47, 6.50)ARNI1.29(−1.96, 4.55)0.78(−5.32, 6.88)SGLT2i1.71(−2.79, 6.22)1.20(−2.73, 5.13)0.42(−4.25, 5.09)RASi3.54(0.60, 6.48)3.02(−2.92, 8.96)2.24(−0.93, 5.42)1.82(−2.63, 6.28)MRA3.75(1.61, 5.89)3.23(−2.36, 8.83)2.46(0.00, 4.91)2.03(−1.94, 6.01)0.21(−1.81, 2.23)Placebo League table representing relative effect estimates for LVEF after decomposing RASi/ARNI 0.51 (−5.47, 6.50) 1.29 (−1.96, 4.55) 0.78 (−5.32, 6.88) 1.71 (−2.79, 6.22) 1.20 (−2.73, 5.13) 0.42 (−4.25, 5.09) 3.54 (0.60, 6.48) 3.02 (−2.92, 8.96) 2.24 (−0.93, 5.42) 1.82 (−2.63, 6.28) 3.75 (1.61, 5.89) 3.23 (−2.36, 8.83) 2.46 (0.00, 4.91) 2.03 (−1.94, 6.01) 0.21 (−1.81, 2.23) Meta-analysis of NT-proBNP Descriptive statistics on NT-proBNP were conducted in the 7 studies analyzed, comprising 4 interventions and 14,426 patients. Network map indicated that comparisons between SGLT2i versus placebo and ACEI versus BB were distinct and could not be combined for network meta-analysis. A common meta-analysis of its individual properties showed that the difference in NT-proBNP between SGLT2i and placebo was statistically significant (MD = −0.09, 95% CI: −0.12, −0.06) (Supplementary Fig. 5), suggesting that SGLT2i lowered NT-proBNP compared to placebo, and there was no significant difference in outcomes between ACEI and BB (Supplementary Fig. 6). Sensitivity analysis The results for LVESV, LVEDV, and LVEF showed high statistical variability. To identify the cause of this variability, sensitivity analyses were conducted by excluding each study one by one. After the Vizzardi-2010 study was excluded, the variability in LVESV and LVEDV significantly decreased without changing the overall results. Unlike other studies, Vizzardi-2010 was noticed to mainly include patients with mild to moderate heart failure (NYHA class I to II). This difference was suspected to be the source of variability. For LVEF, no clear cause of variability was found through sensitivity analyses, and it was speculated that patient comorbidities and baseline medications might be contributing factors (Supplementary Fig. 7). Publication bias A funnel plot and Egger’s test were utilized to assess publication bias for indicators with a significant number of included studies (≥ 10). The results of the funnel plot analysis indicated a central distribution of the literature on the effects of various interventions on LVEF, with a concentration within the funnel and symmetrical distribution. The plot showed bilateral symmetry, implying a minimal risk of publication bias (Supplementary Fig. 8). Discussion This comprehensive network meta-analysis evaluated the therapeutic efficacy of various pharmacological interventions in HFrEF, incorporating data from 22 randomized controlled trials encompassing 16,425 patients. The comparative assessment included RASi, ARNI, BB, MRA, and SGLT2i. Utilizing the SUCRA methodology, we systematically ranked these therapeutic agents to optimize pharmacological management in HFrEF patients. Direct comparative analyses revealed distinct therapeutic profiles: MRA demonstrated superior efficacy in reducing LVESV (MD = −22.10, 95% CI: −31.42 to −12.78). Both ARNI and BB effectively reduced LVESVI, with ARNI showing greater efficacy than BB (MD = −13.78, 95% CI: −25.98 to −1.58; MD = −7.00, 95% CI: −13.91 to −0.09, respectively). SGLT2i exhibited multifaceted benefits, significantly reducing LVMI (MD = −6.57, 95% CI: −11.98 to −1.16) and NT-proBNP levels while improving LVEF. Indirect comparisons further elucidated therapeutic hierarchies: MRA outperformed BB, RASi, and SGLT2i in LVESV reduction. ARNI demonstrated superior efficacy over RASi across multiple parameters, including LVESV, LVESVI, and LVEDVI (MD = −5.80, 95% CI: −9.76 to −1.84). BB showed greater improvement in LVEF compared to MRA (MD = 3.54, 95% CI: 0.60 to 6.48). These findings highlight the differential benefits of ARNI, BB, and MRA in myocardial remodeling. Future investigations should incorporate advanced imaging modalities, particularly cardiac ultrasound and magnetic resonance imaging (MRI), to precisely assess chamber dimensions and systolic function in HFrEF patients. This integrated approach will facilitate the development of personalized therapeutic strategies and optimize clinical decision-making in heart failure management. LVESV represents a critical biomarker for evaluating ventricular remodeling severity, particularly in HFrEF patients, where LVESV changes strongly correlate with disease progression and clinical outcomes. Our findings demonstrate that MRAs significantly improve LVESV in HFrEF patients, outperforming BBs, SGLT2is, and RASis (MD = −22.10, 95% CI: −31.42 to −12.78). The available evidence suggests that MRA-mediated inhibition of myocardial fibrosis and collagen deposition may represent a key mechanism for LVESV improvement in HFrEF patients, potentially optimizing ventricular structure and function. However, these findings derived from indirect comparisons require cautious interpretation due to inherent methodological limitations. Potential confounding factors including inter-study variations in patient demographics (age, comorbidities), therapeutic regimens (dosage, treatment duration), and endpoint definitions may influence result reliability and introduce bias. While ACEIs and angiotensin II receptor blockers (ARBs) form the cornerstone of chronic heart failure management, their long-term efficacy is limited by the “aldosterone escape” phenomenon, where plasma aldosterone levels rebound despite continued therapy. This is particularly relevant given the strong association between elevated aldosterone levels and heart failure severity. MRAs address this limitation by directly inhibiting aldosterone receptor overactivation, thereby mitigating aldosterone-induced myocardial damage and promoting reverse ventricular remodeling. Notably, while BBs demonstrated moderate LVESV reduction, SGLT2is showed no significant effect on this parameter. However, the current evidence base lacks sufficient direct comparative data between MRAs and BBs, and between MRAs and SGLT2is, precluding definitive efficacy rankings. These findings underscore the need for well-designed, large-scale clinical trials to validate these observations and establish robust therapeutic hierarchies in HFrEF management. ARNIs exert their therapeutic effects through dual mechanisms: inhibition of neprilysin-mediated peptide degradation and blockade of angiotensin II receptors. This unique pharmacological action promotes vasodilation, enhances glomerular filtration rate, and increases natriuresis and diuresis, thereby reducing arterial vascular resistance and cardiac afterload. Clinical evidence from the EVALUATE-HF and PROVE-HF trials has demonstrated that ARNIs significantly reduce NT-proBNP levels while improving cardiac structure and function, effectively promoting reverse ventricular remodeling. Our findings corroborate and extend previous research, demonstrating ARNI’s superior efficacy in reversing ventricular remodeling compared to conventional RASis. Specifically, ARNIs showed pronounced benefits in reducing LVESVI (MD = −13.78, 95% CI: −25.98 to −1.58) and LVEDVI (MD = −5.80, 95% CI: −9.76 to −1.84). These results highlight ARNI’s significant and sustained effects on ventricular remodeling in HFrEF patients, establishing it as a superior therapeutic option compared to conventional antihypertensive agents. Interestingly, SUCRA analysis revealed higher rankings for placebo than MRAs in LVESVI and LVEDVI improvement. This unexpected finding may be attributed to two factors: (1) the inclusion of only one MRA study in the mixed-effects model, and (2) substantial loss to follow-up in the original study (11% in the MRA group and 18% in the placebo group). Importantly, the original study provided no evidence of MRA efficacy on LVESVI and LVEDVI parameters, potentially explaining the lack of significant difference between MRA and placebo in our analysis. These findings underscore the need for further well-designed studies to clarify MRA’s role in ventricular remodeling. Clinical trials investigating the therapeutic efficacy of BB in heart failure management have revealed a biphasic response in cardiac function. Initially, BB administration exhibits a transient inhibitory effect on myocardial performance, as indicated by a reduction in LVEF. However, prolonged therapeutic intervention extending beyond three months demonstrates significant functional improvement, manifested by progressive LVEF enhancement. Moreover, extensive clinical evidence suggests that BB therapy effectively attenuates ventricular remodeling processes, reduces chamber dimensions, and improves ventricular geometry, indicating potential reversal or retardation of myocardial remodeling following 4–12 months of continuous treatment. The current investigation corroborates these findings, demonstrating that BB administration significantly reduces LVESV and LVESVI while concurrently improving LVEF, consistent with previous clinical observations. Notably, our comparative analysis revealed superior LVEF improvement with BB therapy relative to MRA treatment. This differential therapeutic effect may be attributable to two key methodological factors: first, 16 of the 22 included studies implemented a follow-up duration exceeding three months; second, the BB treatment cohort demonstrated substantially larger sample size compared to MRA studies. To substantiate these findings and enhance the validity of this indirect comparison, further well-designed randomized controlled trials with extended follow-up periods are warranted. These considerations underscore the importance of meticulous pharmacological selection in clinical practice, emphasizing the need for evidence-based therapeutic decision-making in heart failure management. SGLT2i, originally developed as antidiabetic agents, exhibit pleiotropic cardioprotective effects mediated through multiple pathways [37–40]: hemodynamic regulation, metabolic modulation, ion homeostasis maintenance, and attenuation of myocardial remodeling. The therapeutic benefits primarily stem from their anti-inflammatory, anti-fibrotic, and anti-hypertrophic properties, which collectively ameliorate ventricular remodeling. Preclinical studies [38] and clinical meta-analyses [37] consistently demonstrate significant improvements in ventricular geometry and function, particularly in LVEF and LVMI. Our findings corroborate these observations and further establish the superior efficacy of SGLT2i in LVMI reduction, thereby reinforcing their mechanistic role in cardiac protection. Analysis of NT-proBNP levels demonstrated that the comparative effectiveness between SGLT2i and placebo was independent of that between ACEI and BB. The evidence network mapping revealed an absence of articulation points, precluding the feasibility of conducting a conventional network meta-analysis. Therefore, we performed separate pairwise meta-analyses for each comparative group. The results demonstrated that SGLT2i treatment was associated with a significant reduction in NT-proBNP levels compared to placebo (p < 0.05). In contrast, no statistically significant difference in NT-proBNP reduction was observed between ACEI and BB therapies (p > 0.05), potentially due to the limited evidence base, as only a single clinical study was available for this particular comparison. When considered in conjunction with our previous findings regarding LVMI and LVEF outcomes, these results collectively suggest that SGLT2i therapy exerts multifaceted cardioprotective effects, including significant reduction of NT-proBNP levels, substantial decrease in LVMI, and notable improvement in LVEF. HFrEF remains a significant challenge in clinical treatment as the end stage of various cardiac diseases. Our systematic analysis demonstrates distinct therapeutic advantages of key agents: ARNI and MRA excel in reducing end-systolic volume and reversing ventricular remodeling, while BB shows superior efficacy in enhancing LVEF. These findings support an imaging-guided precision approach: MRA for marked LVESV enlargement, SGLT2i for elevated LVMI, and ARNI replacement of RASi in volume-overloaded patients. This optimized strategy improves clinical outcomes while reducing healthcare expenditures. This strategy not only significantly improves treatment outcomes but also reduces healthcare costs through precision medication. Notably, ARNI outperforms conventional RASi in volumetric parameters, solidifying its role as first-line HFrEF therapy. Similarly, SGLT2i exhibits glycemic-independent cardioprotection, expanding its potential in non-diabetic populations. However, the quadruple regimen carries characteristic risks: RASi/ARNI-associated hypotension/renal impairment, BB-induced conduction abnormalities, MRA-related hyperkalemia, and SGLT2i-linked genitourinary infections. Consequently, vigilant monitoring and personalized dosing are imperative for safe implementation [41, 42]. This study identifies several critical research gaps. First, while focusing on monotherapy for ventricular remodeling, clinical practice necessitates complex combination therapies with varying administration sequences. Future investigations should evaluate optimal drug combinations and sequencing for ventricular remodeling and long-term outcomes. Secondly, the results of this study are primarily based on indirect comparative evidence. Except for the LVESVI indicator, which forms a closed loop, there is a lack of direct comparative data for the remaining outcome indicators. Song et al. [43] confirmed that the reliability of indirect comparisons depends on the homogeneity of the studies. Although this study included high-quality RCTs and there was no significant publication bias, it was still limited by the heterogeneity of population characteristics, intervention protocols, and design parameters, which may affect the reliability of the results. Additionally, the rising global incidence of heart failure in young adults (15–49 years) demands urgent attention [44]. Beyond traditional risk factors (such as obesity, diabetes, hypertension), emerging contributors [44–46] like metabolic syndrome, genetic predisposition, autoimmune disorders, and SARS-CoV-2 infection require thorough investigation through targeted subgroup analyses. Finally, real-world evidence is needed to assess precision medicine strategies’ impact on healthcare utilization and prognosis. These knowledge gaps underscore the necessity for large-scale, multicenter RCTs to validate current findings and strengthen evidence-based therapeutic approaches. Limitation This study has several limitations: (1) it did not consider the variations in drug dosage and duration of administration; (2) it lacked directly comparable results, as the findings were based on indirect comparisons, making it impossible to evaluate the differences between direct and indirect comparisons; (3) the study did not account for baseline medication, focusing solely on the effects of monotherapy on ventricular remodeling. Conclusions Through a combination of traditional meta-analysis and network meta-analysis, we investigated the therapeutic effects of various pharmacological agents, including RASi, ARNI, BB, MRA, and SGLT2i, on myocardial remodeling in patients with chronic heart failure. A comparison of indirect comparison based on network meta-analysis revealed that the MRA exhibited the most significant effect in reducing patients’ LVESV, demonstrating superiority over BB, SGLT2i, and RASi. Meanwhile, SGLT2i was shown to lower LVMI. Furthermore, ARNI also showed a significant advantage over RASi in reducing LVESV, LVESVI, and LVEDVI. In contrast, BB demonstrated superiority over MRA in improving LVEF. Overall, the study provides a significant guideline for the management of pharmacotherapy in patients with HFrEF, which helps clinicians to make more scientific choices when developing a treatment plan. Supplementary Information Supplementary Material 1. Acknowledgements We extend our gratitude to all the authors who contributed to and managed the data for this meta-analysis. Abbreviations Heart failure with a reduced ejection fraction Sodium-glucose cotransporter protein 2 inhibitor Angiotensin receptor-neprilysin inhibitor Angiotensin-converting enzyme inhibitor Angiotensin II receptor blocker Renin-angiotensin system inhibitor Mineralocorticoid receptor antagonist Beta-blocker Left ventricular mass index Left ventricular end-systolic volume Left ventricular end-diastolic volume Left ventricular end-systolic volume index Left ventricular end-diastolic volume index Left ventricular ejection fraction N-terminal B-type natriuretic peptide Mean difference Confidence interval Surface under the cumulative ranking curve Authors’ contributions All authors contributed to the study’s conception and design. Material preparation, data collection, and analysis were performed by X.W. Wang, H. Li, and G.Q. Wang. The initial draft of the manuscript was prepared by X.W. Wang, while Z.H. Wang was responsible for the investigation. W.Z. provided supervi-sion throughout the research process. X.W. Wang and Y.N. Sheng contributed to the validation of the study. Data visualization was carried out by H.P. and H.L. The manuscript was critically reviewed and ed-ited by M.Z. and Y.Y. Shang. Funding This work was supported by research grants from the National Natural Science Foundation of China (82070392, 81900332), the Natural Science Foundation of Shandong Province (ZR2021QH033), Taishan Scholars (No. tsqn202103146) and the Clinical Research Center of Shandong University (No. 2020SDUCRCC027). Data availability The data and materials utilized to support the conclusions of this meta-analysis are available from the corresponding author upon reasonable and justified request. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Xi-Wen Wang is the first author of this article. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Yuan-Yuan Shang, Email: shang@qiluhospital.com. Ming Zhong, Email: zhongming2sdu@163.com. References Associated Data Supplementary Materials Supplementary Material 1. Data Availability Statement The data and materials utilized to support the conclusions of this meta-analysis are available from the corresponding author upon reasonable and justified request.