Survival benefit of guideline-directed medical therapy in older patients with heart failure: a systematic review and meta-analysis

Abstract

Background

Heart failure-related mortality in older patients has increased. Guideline-directed medical therapy (GDMT) in heart failure with reduced ejection fraction (HFrEF) remains insufficient in this population. This study assessed the survival benefit of GDMT in elderly patients aged ≥65 years old with HFrEF.

Methods

MEDLINE, EMBASE and Cochrane Library were searched from inception to 5 February 2025. Inclusion criteria were age ≥65 years old with HFrEF, GDMT vs no GDMT and reported survival data. Pooled prevalence, HR and risk ratio (RR) with 95% CI were analysed. Heterogeneity was further explored through subgroup analyses. We evaluated potential small-study effects using Begg’s rank correlation test, Egger’s regression test, funnel plot inspection and trim-and-fill analysis. Publication bias was analysed visually using a funnel plot.

Results

34 studies with 92 916 patients from 1994 to 2024 were included. The GDMT group received three of the four main drug classes in three studies, two drug classes in four studies and at least one class in the remaining studies. There were no reported data on the survival benefits of older patients receiving four classes of GDMT. Frail older patients were less likely to receive GDMT (RR 0.63 (0.45–0.89), I2 50%, p=0.009). There was a survival benefit of GDMT in older patients with left ventricular (LV) systolic dysfunction (HR 0.69 (0.64–0.74) p<0.001, Tau² 0.03, I² 71%). Older patients receiving beta-blockers were more likely to develop bradycardia (RR 3.98 (2.72–5.82), p<0.001, I2 0%). There was no difference in hypotension, acute renal impairment and hypoglycaemia between the GDMT and control group.

Conclusions

The use of GDMT in older patients with LV systolic dysfunction was associated with a survival benefit. However, the presence of frailty may limit the successful implementation and tolerability of these therapies. High-quality studies are urgently needed to guide the development of novel strategies aimed at optimising the management of older, frail patients with LV systolic function.

Introduction

Heart failure-related mortality in older population has increased since 2012.¹ Compared with the younger population, the older group carried a higher incidence of heart failure-related events, exacerbations and hospital admissions.² Heart failure-related mortality and morbidity in the older population confers a substantial burden to the healthcare system.³ This is a matter of increasing concern as the population with heart failure (HF) aged 65 and above is projected to reach 95 million in 2060.⁴

Guideline-directed medical therapy (GDMT) is a cornerstone to reduce mortality and morbidity in patients with HF and reduced left ventricular ejection fraction (HFrEF). The 2021 European Society of Cardiology (ESC) and the 2022 American Heart Association (AHA) guideline recommends initiating four main drug classes, namely, (i) one of an angiotensin-converting enzyme inhibitor (ACEi), angiotensin receptor blocker (ARB) or angiotensin receptor-neprilysin inhibitor (ARNi); (ii) beta-blockers (BB); (iii) mineralocorticoid receptor antagonists (MRA); and (iv) sodium-glucose cotransporter-2 inhibitors (SGLT2i).^{5 6} Each of the four GDMT classes has a life-saving effect with an additive benefit.

The challenges to GDMT implementation in older adults with HF are bounded by concerns of age-associated physiologic alterations as well as pharmacodynamic and pharmacokinetic variations that can make them more susceptible to adverse events. Frailty is common in older adults and can worsen HF prognosis; conversely, HF can predispose individuals to developing frailty, which increases the risk of HF progression and prognosis. The evidence on the utility of GDMT in this population remains insufficient, leading to suboptimal initiation and maintenance of treatment in this vulnerable group.⁷

The objective of this study was to assess the survival benefit of GDMT in older patients aged ≥65 years old with HFrEF.

Methods

Search strategy

This review was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines (online supplemental appendix 1).

Data source and study search

We searched three electronic data sources (MEDLINE, EMBASE and the Cochrane Library) from inception to 5 February 2025 using a combination of MeSH headings and keywords (and their derivatives): beta-blockers (BB), Renin-angiotensin system inhibitor (RASi) which include angiotensin-converting enzyme inhibitor (ACEi), angiotensin II receptor blocker (ARBs), Angiotensin receptor-neprilysin inhibitor (ARNi), mineralocorticoid receptor antagonist (MRA), sodium-glucose cotransporter-2 inhibitor (SGLT2i), heart failure, ventricular dysfunction, elderly and older. The keyword strategy used is presented in online supplemental appendix 2. Reference lists of relevant studies and review articles are searched manually for potentially eligible studies not captured in the primary search.

Articles were included if they met all of the following criteria (i) randomised controlled trials (RCTs), non-randomised prospective studies (case control or controlled cohort) and observational studies; (ii) study population that included patients ≥65 years old or included subgroup analysis of patients aged ≥65 years, consistent with geriatric prescribing frameworks,⁸ cardiovascular risk scores (Thrombolysis In Myocardial Infarction risk (TIMI)⁹ risk score for unstable angina or Non-ST Elevated myocardial infarction, dual anti-platelet therapy (DAPT) predictive ischaemic and bleeding risk scores after percutaneous coronary intervention,¹⁰ CHADsVASc¹¹ ischaemic stroke risk score for non valvular atrial fibrillation and guideline recommendations;¹² (iii) patients with left ventricular systolic dysfunction (left ventricular ejection fraction (LVEF) <50%); and (iv) outcomes that compared the treatment effect between GDMT (medications from at least one of more class) and no GDMT (defined as the placebo group in RCT and the group with no active treatment in observational studies) groups with HR or survival curve. No language restrictions were applied. The primary endpoint was overall survival (OS). Secondary outcomes were clinical characteristics of older patients with left ventricular systolic dysfunction.

The search of electronic databases was conducted by two independent reviewers (E.T. and C.M.W.) who identified potentially relevant articles based on title or abstract. Full-text articles were then retrieved for further assessment to determine whether they met the inclusion criteria. Discrepancies were resolved by consensus. Emails were sent to first authors of the primary studies requesting additional data. There was no involvement of patients or public involvement in the design, conduct, analysis or reporting in this meta-analysis.

Quality assessment

The Risk of Bias 2 (RoB2) assessment tool¹³ was used to assess the quality of RCTs and Newcastle-Ottawa scale (NOS)¹⁴ for observational studies. The NOS scores of 0–3, 4–6 and 7–9 are interpreted as having high, moderate or low risk of bias (RoB), respectively.¹⁵ The certainty of evidence for the primary outcome was assessed using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) framework which incorporates study quality (risk of bias), inconsistency (comparison of effect estimates across studies), indirectness (applicability of the population, intervention, comparator and outcomes to the clinical decision), imprecision (certainty of CI) and probability of publication bias. The overall quality of evidence was categorised as high, moderate, low and very low by combining the aforementioned five parameters.¹⁶ Scoring was performed by two independent authors and differences were resolved by consensus.

Data analysis

Two authors (E.T. and M.W.C.) independently extracted predefined data. Extracted information included study characteristics (author, geographic location, design, sample size) and patient characteristics (age, sex, body mass index, current smoker, New York Heart Association (NYHA) Class III–IV, LVEF %, ischaemic heart disease, atrial fibrillation (AF), hypertension, hyperlipidaemia, diabetes, stroke/transient ischaemic attack (TIA), chronic obstructive pulmonary disease, dementia, dialysis-dependent renal failure), the prevalence of each pillar of GDMT, adverse events and the extraction of hazards ratio or survival curves. Authors of original articles responded to our query with clarification and/or assisted in reanalysis with a response rate of 3%. Consequently, analyses were conducted using available data applying sensitivity analyses, and no imputation of missing secondary outcome data was undertaken.

Statistical analysis

Pooled prevalence was performed with MetaXL (V.5.3; EpiGear International Pty Ltd).

The random-effects model (DerSimonian and Laird method) was applied to estimate the pooled prevalence and 95% CI.¹⁷ To account for extreme prevalence data, prevalence estimates were transformed with the use of the double arcsine method and then back-transformed for ease of interpretation.¹⁸

Pooled hazards ratio, OR and weighted mean difference were calculated using Review Manager (RevMan; version 5.4; The Cochrane Collaboration). For categorical variables, pooled risk ratio (RR) and associated 95% CIs were calculated with the use of a Mantel-Haenszel model; for continuous variables, mean differences and associated 95% CIs were calculated with an inverse-variance model. Pooled survival data was analysed using HR with 95% CI to compare the survival statistics between older HFrEF patients receiving GDMT versus no GDMT. For studies that only reported survival curves, Log (HR) and SE were estimated from the available survival curves using a spreadsheet developed by Tierney et al.¹⁹

Heterogeneity was assessed using the Higgins I2 statistic and the χ2 statistic to estimate the total variation across studies where statistical heterogeneity is defined as I2 >50% and/or χ2 statistic p < 0.10. Heterogeneity was further explored through subgroup analyses with the use of the following categories of study characteristics: (1) age cut-off (age ≥65 years vs age ≥70 vs age ≥80 years), (2) number of GDMT class (at least three vs at least two vs only one), (3) year of evolution of HF treatment (before 2001 vs 2001 to 2020 vs at and after 2021) (RAASi/BB were core treatment recommendations for the HF treatment in the 1990s, MRA was incorporated in 2001 by both the ESC²⁰ and AHA²¹ and SGLT2 inhibitors more recently integrated—first in the 2021 ESC²² and subsequently in the 2022²³ AHA update), (4) study type (RCT vs observational study), (5) risk of bias (high risk vs low to moderate risk), (6) LVEF cut-off (EF ≤50% vs EF ≤40% vs EF ≤35% vs HFrEF cut-off not reported), (7) sample size (≥ 1000 versus <1000 patients) and (8) geographical location (American centres vs European centres vs Asian centres vs Multinational studies). Meta-regression was conducted with study-level age cut-off and the number of GDMT agents, used as continuous moderators, to assess potential effect modification. Sensitivity analyses were performed to explore the impact of study quality on pooled estimates. We evaluated potential small-study effects using Begg’s rank correlation test, Egger’s regression test, funnel plot inspection and trim-and-fill analysis. Publication bias was analysed visually using a funnel plot. We performed a sensitivity analysis to examine the robustness of the results, systematically removing one study from the analysis and recalculating the results.

Results

Study selection

A total of 5794 articles were retrieved with the use of the search strategy. After screening was performed by abstract and title, 528 articles were selected for full text assessment. 34 studies met the inclusion criteria and were included in this systematic review and meta-analysis (figure 1).

Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) 2020 flow diagram for updated systematic reviews which included searches of databases, registers and other sources. GDMT, guideline-directed medical therapy; HFpEF, heart failure with preserved ejection fraction.

Study characteristics

We included 34 studies^24–57 [dataset]²⁴ Abe T, Jujo K, Kametani M, et al Data from: Prognostic impact of additional mineralocorticoid receptor antagonists in octogenarian heart failure patients. ESC heart failure 2020. https://pubmed.ncbi.nlm.nih.gov/32860346/ [dataset]²⁵ Abe T, Jujo K, Maeda D, et al Data from: The interaction between physical frailty and prognostic impact of heart failure medication in elderly patients. ESC heart failure 2023. https://pubmed.ncbi.nlm.nih.gov/36824014/ [dataset]²⁶ Ahmed A, Centor RM, Weaver MT, et al Data from: A propensity score analysis of the impact of angiotensin-converting enzyme inhibitors on long-term survival of older adults with heart failure and perceived contraindications. American Heart Journal 2005 https://pubmed.ncbi.nlm.nih.gov/15990761/ [dataset]²⁷ Bayoumi E, Lam PH, Dooley DJ, et al Data from: Spironolactone and Outcomes in Older Patients with Heart Failure and Reduced Ejection Fraction. The American journal of medicine 2019. https://pmc.ncbi.nlm.nih.gov/articles/PMC6511886/ [dataset]²⁸ Deedwania PC, Gottlieb S, Ghali JK, et al Data from: Efficacy, safety and tolerability of beta-adrenergic blockade with metoprolol CR/XL in elderly patients with heart failure. European heart journal 2004 https://pubmed.ncbi.nlm.nih.gov/15288157/²⁹ Del Sindaco D, Pulignano G, Cioffi G, et al Data from: Safety and efficacy of carvedilol in very elderly diabetic patients with heart failure. Journal of cardiovascular medicine (Hagerstown, Md) 2007. https://pubmed.ncbi.nlm.nih.gov/17700396/³⁰ Filippatos G, Anker SD, Butler J, et al Data from: Effects of empagliflozin on cardiovascular and renal outcomes in heart failure with reduced ejection fraction according to age: a secondary analysis of EMPEROR-Reduced. European journal of heart failure 2022. https://pubmed.ncbi.nlm.nih.gov/36194680/³¹ Flather MD, Shibata MC, Coats AJS, et al Data from: FASTTRACK Randomized trial to determine the effect of nebivolol on mortality and cardiovascular hospital admission in elderly patients with heart failure (SENIORS). European Heart Journal 2005 https://pubmed.ncbi.nlm.nih.gov/15642700/ [dataset]³² Franco Pelaez JA, Cortes Garcia M, Romero Daza AM, et al Data from: Relationship between different doses of beta-blockers and prognosis in elderly patients with reduced ejection fraction. International journal of cardiology 2016. https://pubmed.ncbi.nlm.nih.gov/27389445/ [dataset]³³ Gilstrap L, Solomon N, Chiswell K, et al Data from: The Association Between Beta-blocker and Renin-Angiotensin System Inhibitor Use After Heart Failure With Reduced Ejection Fraction. Hospitalization and Outcomes in Older Patients. Journal of Cardiac Failure 2023 https://pubmed.ncbi.nlm.nih.gov/36516937/ [dataset]³⁴ Khoury J, Ghersin I, Braun E, et al Data from: Adherence to Guidelines in Heart Failure, Is It Valid for Elderly Patients? The Israel Medical Association journal : IMAJ 2022. https://pubmed.ncbi.nlm.nih.gov/36436045/ [dataset]³⁵ Martinez FA, Serenelli M, Nicolau JC, et al Data from: Efficacy and Safety of Dapagliflozin in Heart Failure with Reduced Ejection Fraction According to Age: Insights from DAPA-HF. Circulation 2020. https://pubmed.ncbi.nlm.nih.gov/31736328/ [dataset]³⁶ Martinez-Milla J, Garcia MC, Urquia MT, et al Data from: Blockade of Renin-Angiotensin-Aldosterone System in Elderly Patients with Heart Failure and Chronic Kidney Disease: Results of a Single-Center, Observational Cohort Study. Drugs & aging 2019. https://pmc.ncbi.nlm.nih.gov/articles/PMC7868915/ (dataset)³⁷ Molnar AO, Petrcich W, Weir MA, et al Data from: The association of beta-blocker use with mortality in elderly patients with congestive heart failure and advanced chronic kidney disease. Nephrology, dialysis, transplantation: official publication of the European Dialysis and Transplant Association - European Renal Association 2020. https://pubmed.ncbi.nlm.nih.gov/31495887/ [dataset]³⁸ Packer M, Coats AJ, Fowler MB, et al Data from: Carvedilol Prospective Randomized Cumulative Survival Study Group. Effect of carvedilol on survival in severe chronic heart failure. N Engl J Med 2001 https://pubmed.ncbi.nlm.nih.gov/11386263/ [dataset]³⁹ Pascual-Figal DA, Redondo B, Caro C, et al Data from: Comparison of late mortality in hospitalized patients>70 years of age with systolic heart failure receiving beta blockers vs those not receiving beta blockers. The American journal of cardiology 2008. https://pubmed.ncbi.nlm.nih.gov/19064029/ [dataset]⁴⁰ Pedone C, Pahor M, Carosella L, et al Data from: Use of angiotensin-converting enzyme inhibitors in elderly people with heart failure: prevalence and outcomes. The journals of gerontology Series A, Biological sciences and medical sciences 2004. https://pubmed.ncbi.nlm.nih.gov/15304537/ [dataset]⁴¹ Pitt B, Zannad F, Remme WJ, et al Data from: The effect of spironolactone on morbidity and mortality in patients with severe heart failure. Randomized Aldactone Evaluation Study Investigators. N Engl J Med 1999. https://pubmed.ncbi.nlm.nih.gov/10471456/ [dataset]⁴² Rochon PA, Sykora K, Bronskill SE, et al Data from: Use of angiotensin-converting enzyme inhibitor therapy and dose-related outcomes in older adults with new heart failure in the community. Journal of General Internal Medicine 2004. https://pubmed.ncbi.nlm.nih.gov/15209607/ [dataset]⁴³ Sanam K, Bhatia V, Parvataneni S, et al Data from: Discharge initiation of ACE inhibitors or ARBs is associated with significantly lower 30-day all-cause readmission in hospitalized older patients with heart failure and reduced ejection fraction. Journal of the American College of Cardiology 2014 https://www.jacc.org/doi/abs/10.1016/s0735–1097%2814%2960562–3 [dataset]⁴⁴ Sargento L, Simoes AV, Longo S, et al Data from: Treatment with Optimal Dose Angiotensin-Converting Enzyme Inhibitors/Angiotensin Receptor Blockers Has a Positive Effect on Long-Term Survival in Older Individuals (Aged >70 Years) and Octogenarians with Systolic Heart Failure. Drugs & aging 2016. https://pubmed.ncbi.nlm.nih.gov/27568454/ [dataset]⁴⁵ Seo WW, Park JJ, Park HA, et al Data from: Guideline-directed medical therapy in elderly patients with heart failure with reduced ejection fraction: A cohort study. BMJ Open 2020 https://pubmed.ncbi.nlm.nih.gov/32034017/ [dataset]⁴⁶ Shlipak MG, Browner WS, Noguchi H, et al Data from: Comparison of the effects of angiotensin converting-enzyme inhibitors and beta blockers on survival in elderly patients with reduced left ventricular function after myocardial infarction. The American journal of medicine 2001. https://pubmed.ncbi.nlm.nih.gov/11331052/ [dataset]⁴⁷ Silva JE, Melo N, Ferreira AI, et al Data from: Prognostic impact of neurohormonal modulation in very old patients with chronic heart failure. Age and ageing 2022. https://pubmed.ncbi.nlm.nih.gov/35363254/ [dataset]⁴⁸ Stolfo D, Uijl A, Benson L, et al Data from: Association between beta-blocker use and mortality/morbidity in older patients with heart failure with reduced ejection fraction. A propensity score-matched analysis from the Swedish Heart Failure Registry. European journal of heart failure 2020 https://pubmed.ncbi.nlm.nih.gov/31478583/ [dataset]⁴⁹ Sung SH, Wang TJ, Cheng HM, et al Data from: Clinical Characteristics and Outcomes in the Very Elderly Patients Hospitalized for Acute Heart Failure: Importance of Pharmacologic Guideline Adherence. Scientific reports 2018. https://pubmed.ncbi.nlm.nih.gov/30250052/ [dataset]⁵⁰ Urrutia A, Lupon J, Altimir S, et al Data from: Use of betablockers in elderly patients with congestive heart failure. Uso de bloqueadores beta en pacientes ancianos con insuficiencia cardiaca 2006. https://www.elsevier.es/es-revista-medicina-clinica-2-articulo-uso-bloqueadores-beta-pacientes-ancianos-insuficiencia-cardiaca-13084869 [dataset]⁵¹ Zannad F, McMurray JJ, Krum H, et al Data from: Eplerenone in patients with systolic heart failure and mild symptoms. N Engl J Med 2011 https://www.nejm.org/doi/full/10.1056/NEJMoa1009492 [dataset]⁵² Araújo C, Marco I, Restrepo-Córdoba M, et al Data from: An Observational Study of Evidence-Based Therapies in Older Patients with Heart Failure with Reduced Ejection Fraction: Insights from a Dedicated Heart Failure. Clinic J Clin Med 2024 https://pubmed.ncbi.nlm.nih.gov/39685630/ [dataset]⁵³ Germán J, García M, López C, et al Data from: Impact of SGLT2 Inhibitors on Very Elderly Population with Heart Failure with Reduce Ejection Fraction: Real Life Data. Biomedicines 2024. https://pubmed.ncbi.nlm.nih.gov/39062080/ [dataset]⁵⁴ Huang C, Park J, Liu I, et al Data from: Effectiveness and safety of early treatment with spironolactone for new-onset acute heart failure. J Hosp Med 2024 https://pubmed.ncbi.nlm.nih.gov/38415888/ [dataset]⁵⁵ Monzo L, Girerd N, Duarte K, et al Data from: Time to clinical benefit of eplerenone among patients with heart failure and reduced ejection fraction: A subgroups analysis from the EMPHASIS-HF trial. Eur J Heart Fail 2023 https://pubmed.ncbi.nlm.nih.gov/37370197/ [dataset]⁵⁶ Roca L, García M, Germán J, et al Data from: Use and Benefit of Sacubitril/Valsartan in Elderly Patients with Heart Failure with Reduced Ejection Fraction. J Clin Med 2024. https://pubmed.ncbi.nlm.nih.gov/39200914/ (dataset)⁵⁷ Svanström H, Mkoma G, Hviid A, et al Data from: SGLT-2 inhibitors and mortality among patients with heart failure with reduced ejection fraction: linked database study. BMJ 2024 https://pubmed.ncbi.nlm.nih.gov/39505389/ with 92,916 patients from 1994 to 2024. There were 19 observational retrospective studies,^{24–27 32–34 36 37 42 43 46 47 49 50 52–54 57} seven prospective studies and^{29 39 40 44 45 48 56} eight RCTs.^{28 30 31 35 38 41 51 55} All eight RCTs^{28 30 31 35 38 41 51 55} included in this review were determined as low risk of bias across the assessed domains. In contrast, among the 26 observational studies,^{24–27 29 32–34 36 37 39 40 42–50 52–54 56 57} three studies^{25 40 50} were rated as high RoB and the remaining observational studies were low or moderate risk^{24 26 27 29 32–34 36 37 39 42–49 52–54 56 57} (online supplemental appendix 3).

The older group comprised patients in studies defined over the ages of 65 (14 studies^{25 26 28 33 37 38 40–43 45 46 54 55}), 70 (five studies^{29 31 39 44 57}), 75 (ten studies^{27 30 32 35 36 50–53 56}) and 80 (five studies^{24 34 47–49}). The effect of GDMT vs placebo or no active treatment was assessed with the GDMT group receiving three of the four main drug classes in three studies,^{24 34 49} two classes in four studies^{25 33 45 46} and one class in the remaining included studies.^{26–32 35–44 47 48 50–57} Study characteristics are presented in table 1.

Table 1. Characteristics of included studies of older patients with abnormal left ventricular systolic function.

No	Author	Country	Sample source	Study type	Time-frame	Ejection fraction	Age cut-off (years)	GDMT class
1	Abe24	Japan	Single CV centre – not otherwise reported	Retrospective observational study	2013–2017	<50%	≥80	BB+RASi + MRA
2	Abe et al25	Japan	FRAGILE-HF – multicentre observational cohort and Kitasato University Hospital	Retrospective observational cohort study	2016–2018 2008–2017	<50%	≥65	BB+RASi
3	Ahmed et al26	USA 11 Alabama hospitals	Alabama Quality Assurance Foundation	Cross sectional observational study	1994	<40%	≥65	RASi
4	Araújo et al52	Spain	Hospital Clinico San Carlos	Retrospective observational studies	2022–2023	<40%	≥75	SGLT2i
5	Bayoumi	USA 259 hospitals in 48 states	OPTIMIZE-HF registry	Retrospective observational propensity score-matched cohorts	2003–2004	≤35%	≥75	MRA
6	Deedwania	14 countries	MERIT-HF – secondary analysis	Prospective placebo-controlled double-blind clinical trial - Subgroup analysis	1997–1998	≤40%	≥65	BB
7	Del Sindaco et al29	Italy	Heart failure clinics	Prospective observational study	2001	<40%	≥70	BB
8	Filippatos	Multinational	EMPEROR-Reduced – secondary analysis	Randomised controlled trial - Subgroup analysis	2017–2019	≤40%	≥75	SGLT2i
9	Flather et al31	UK	SENIORS	Randomised double blind multicentre international trial	2000–2002	≤35%	≥70	BB
10	Franco Peláez et al32	Spain	Hospital Universitario Fundacion Jimenez Diaz	Observational cohort study	2008–2014	≤35%	≥75	BB
11	Balaguer Germán et al53	Spain	Hospital Universitario Fundacion Jimenez Diaz	Retrospective observational study	2019–2022	<40%	≥75	SGLT2i
12	Gilstrap et al33	USA	GWTG-HF	Retrospective observational study	2008–2016	HFrEF	≥65	BB±RASi
13	Huang et al54	USA 215 sites in California	MIDTERM-HF Cohort	Retrospective cohort study	2016–2021	≤40%	≥65	MRA
14	Khoury et al34	Israel	Rambam Health Care	Retrospective observational study	2008–2018	HFrEF	≥80	BB±RASi +/-MRA
15	Martinez et al35	410 centres, 20 countries	DAPA-HF – secondary analysis	Randomised placebo controlled trial – subgroup analysis	2017–2018	≤40%	≥75	SGLT2i
16	Martínez-Milla et al36	Spain	Hospital Fundacion Jimenez Diaz	Single centre observational cohort study	2008–2014	≤35%	≥75	RASi
17	Molnar et al37	Canada	Institute for Clinical Evaluative Sciences (ICES) in Ontario	Matched retrospective cohort study	2002–2014	NR	>65	BB
18	Monzo et al55	278 centres in 29 countries	EMPHASIS-HF trial – subgroup analysis	Multicentre randomised double-blind clinical trial – subgroup analysis	2006–2010	≤35%	≥65	MRA
19	Packer et al38	334 centres in 21 countries	COPERNICUS trial – secondary analysis	Randomised double-blind trial - subgroup analysis	1997–2000	≤25%	≥65	BB
20	Pascual-Figal et al39	Spain	Cardiology department of a tertiary university hospital	Prospective observational study	2002–2003	<50%	≥70	BB
21	Pedone et al40	Italy	Teaching and non-teaching hospitals in Italy	Prospective observational study	1998	NR	≥65	RASi
22	Pitt et al41	195 centres in 15 countries	RALES trial – secondary analysis	Randomised double blind study - subgroup analysis	1995–1996	≤35%	≥67	MRA
23	Roca et al56	Spain	Hospital Universitario Fundacion Jimenez Diaz	Single centre prospective observational cohort study	2016–2020	<40%	≥75	ARNi
24	Rochon et al42	Canada	Canadian Institute for Health Information database	Observational cohort study	1998–1999	NR	>65	RASi
25	Sanam et al43	USA 106 Alabama hospitals	Alabama Heart Failure project	Retrospective observational study	1998–2001	<45%	≥65	RASI
26	Sargento et al44	Portugal	Outpatient HF clinic in a university hospital	Prospectively recruited population study	2010–2012	<40%	≥70	RASi
27	Svanström et al57	Denmark	Danish Heart Failure Registry – subgroup analysis	Linked database study	2020–2023	≤40%	≥70	SGLT2i
28	Seo et al45	Korea	KorAHF registry	Prospective cohort study	2011–2014	≤40%	≥65	BB+RASi
29	Shlipak et al46	USA	Medicare beneficiaries	Comparative observational study	1994–1995	≤40%	≥65	BB+RASi
30	Silva et al47	Portugal	Tertiary care hospital in Porto, Portugal	Retrospective observational study	2012–2018	<50%	≥80	BB
31	Stolfo et al48	Sweden	Swede HF	Prospective propensity score-matched cohort study	2000–2015	<40%	≥80	BB
32	Sung et al49	Taipei	HARVEST registry	Observational cohort study	2003–2012	<50%	≥80	BB+RASi + MRA
33	Urrutia	Spain	Hospital Universitari Germans Trias I Pujol	Observational study	2001–2004	NR	≥75	BB
34	Zannad et al51	278 centres, 29 countries	EMPHASIS-HF	Randomised double blind trial – subgroup analysis	2006–2010	<35%	≥75	MRA

ARNi, angiotensin receptor-neprilysin inhibitor; BB, beta-blocker; CV, cardiovascular; d, days; eGFR, estimated glomerular filtration rate; GDMT, guideline-directed medical therapy; HF, heart failure; MRA, mineralocorticoid receptor antagonist; mth, months; no GDMT, <1 of the four GDMT; NR, not reported; NYHA, New York Heart Association; RASi, renin-angiotensin system inhibitor.

Clinical characteristics

We analysed 31 variables that include demographics, comorbidities, cardiac history, medical therapy and adverse events (online supplemental appendices 4-8).

More than half of the older patients with HFrEF were male (pooled prevalence 60.6%, 95% CI 57.6 to 63.6, I² 99%). Comorbidity burden was extracted from each individual study and presented in online supplemental appendix 6. Common comorbidities included hypertension (pooled prevalence 70.2%, 95% CI 67.2 to 73.0, I² 98%), coronary artery disease (pooled prevalence 45.2%, 95% CI 36.3 to 54.3, I² 100%), dyslipidaemia (pooled prevalence 48.2%, 95% CI 44.1 to 52.3, I² 92%), AF (pooled prevalence 40.1%, 95% CI 31.1 to 49.4, I² 100%) and diabetes (pooled prevalence 35.4%, 95% CI 32.3 to 38.6, I² 98%) (table 2, online supplemental appendix 9).

Table 2. Summary of pooled prevalence of demographic, comorbidities, cardiac-related comorbidities and evidence-based medical therapy in elderly patients with heart failure with reduced ejection fraction.

Variables	Studies	Total sample size (n)	Patients (n)	Pooled (prevalence (%))	95% CI (upper to lower)	I2% (p value)
Male	24	96 665	54 385	60.6%	57.67 to 63.57	99% (p<0.001)
Frailty*	3	50 338	6250	7.0%	2.00 to 14.00	98% (p<0.001)
Smoker	13	62 422	5925	15.4%	10.32 to 21.26	99% (p<0.001)
Hypertension	21	86 288	60 429	70.2%	67.23 to 73.03	98% (p<0.001)
Dyslipidaemia	11	53 016	26 869	48.2%	44.17 to 52.28	92% (p<0.001)
Diabetes	20	88 144	33 076	35.4%	32.25 to 38.67	98% (p<0.001)
Stroke/TIA	11	70 390	9798	15.0%	10.42 to 20.23	99% (p<0.001)
Dementia	7	78 538	7024	10.8%	6.30 to 16.39	100% (p<0.001)
Dialysis-dependent	4	49 650	2356	3.4%	1.86 to 5.25	79% (p<0.001)
NYHA class III/IV	14	15 647	7163	33.8%	22.4 to 46.15	100% (p<0.001)
Coronary artery disease	15	85 841	47 273	45.2%	36.32 to 54.31	100% (p<0.001)
Valvular heart disease	3	6944	2591	17.3%	1.21 to 42.75	99% (p<0.001)
Atrial fibrillation	17	77 772	30 025	40.1%	31.09 to 49.38	100% (p<0.001)
Ischaemic cardiomyopathy	10	4837	2839	48.9%	38.06 to 59.86	98% (p<0.001)
RASi	24	97 688	61 800	74.2%	66.62 to 81.08	100% (p<0.001)
BB	23	98 323	69 745	65.1%	50.31 to 78.57	100% (p<0.001)
MRA	20	77 431	9418	33.6%	23.49 to 44.47	100% (p<0.001)
SGLT2i	5	3301	726	16.4%	0.00 to 41.55	100% (p<0.001)
Diuretics	19	38 453	30 054	85.4%	79.76 to 90.18	100% (p<0.001)

^*Frailty defined as hospital frailty risk score ≥5, nursing home residents and any degree of functional disability.

BB, beta-blocker; COPD, chronic obstructive pulmonary disease; GDMT, guideline-directed medical therapy; MRA, mineralocorticoid receptor antagonist; NYHA, New York Heart Association; RASi, renin-angiotensin system inhibitor; SGLT2i, sodium-glucose cotransporter-2 inhibitor; TIA, transient ischaemic attack.

More than 5% of the older HFrEF patients was frail (defined by a hospital frailty risk score ≥5 or nursing home residents) with a pooled prevalence of 7% (95% CI 2.0 to 14.0, I² 98%) (table 2, online supplemental appendix 9). Older patients who were frail were less likely to receive GDMT (RR 0.63, 95% CI 0.45 to 0.89, I² 50%, Tau² 0.05, p=0.009) (table 3, online supplemental appendix 10).

Table 3. Summary of the quantitative pooled effect size of each demographic and comorbid variables in the GDMT versus no GDMT group.

Variables	GDMT		No GDMT		RR (95% CI upper - lower) MD (95% CI upper - lower)	P value	Heterogeneity I2% (p value)
	Total, no.	Events, no. (%)	Total, no.	Events, no. (%)
Clinical characteristics
Age (years)	70 179	77.8 years†	24 444	79.3 years†	MD −1.53 (−2.08 to 0.98)	p<0.001*	93% (p<0.001)
Male	67 087	39 335 (58.6%)	17 704	9692 (54.7%)	RR 0.98 (0.95 to 1.02)	p=0.30	56% (p=0.002)
BMI (kg/m2)	8653	24.2 kg/m2†	2830	23.9 kg/m2	MD 0.34 (−0.13 to 0.80)	p=0.15	82% (p<0.001)
Current smoker	56 381	6025 (10.7%)	4951	493 (10.0%)	RR 1.16 (0.94 to 1.43)	p=0.18	73% (p<0.001)
Frailty	48 198	5908 (12.3%)	2140	342 (16.0%)	RR 0.63 (0.45 to 0.89)	p=0.009*	50% (p=0.13)
NYHA III/IV classification	10 262	4933 (48.1%)	4188	1999 (47.7%)	RR 0.83 (0.74 to 0.94)	p=0.002*	82% (p<0.001)
LVEF (%)	51 240	30.6 %†	4166	31.2 %†	MD −0.60 (−1.96 to 0.77)	p=0.39	92% (p<0.001)
Coronary artery disease	66 909	39 320 (58.8%)	17 799	7704 (43.3%)	RR 1.07 (1.00 to 1.15)	p=0.04*	86% (p<0.001)
Atrial fibrillation	64 067	24 964 (39.0%)	12 035	3822 (31.8%)	RR 0.91 (0.81 to 1.03)	p=0.12	91% (p<0.001)
Ischaemic cardiomyopathy	2195	1349 (61.5%)	1543	1002 (64.9%)	RR 1.11 (0.95 to 1.30)	p=0.002*	71% (p=0.19)
Hypertension	66 817	47 447 (71.0%)	17 237	11 181 (64.9%)	RR 1.07 (1.02 to 1.12)	p=0.01*	86% (p<0.001)
Dyslipidaemia	48 945	24 914 (50.9%)	2981	1319 (44.2%)	RR 1.11 (1.03 to 1.20)	p=0.007*	27% (p 0.22)
Diabetes	67 727	25 224 (37.2%)	18 183	6998 (38.5%)	RR 1.07 (1.00 to 1.16)	p=0.06	79% (p<0.001)
Stroke	54 714	8720 (15.9%)	9282	895 (9.64%)	RR 1.13 (0.57 to 2.23)	p=0.73	98% (p<0.001)
Dementia	63 473	5093 (8.0%)	15 065	1931 (12.8%)	RR 0.78 (0.63 to 0.97)	p=0.03*	86% (p<0.001)
Dialysis	47 453	2223 (4.7%)	1581	120 (7.6%)	RR 0.59 (0.49 to 0.70)	p<0.001*	0% (p=0.62)
Adverse events	GDMT		Placebo
Hypotension  (beta-blocker studies)	2057	142 (6.9%)	2053	131 (6.4%)	RR 2.98 (0.21 to 43.23)	p=0.42	73% (p=0.05)
Bradycardia  (beta-blocker studies)	2057	128 (6.2%)	2053	32 (1.5%)	RR 3.96 (2.72 to 5.82)	p<0.001*	0% (p=0.41)
Acute renal impairment  (SGLT2i studies)	1092	90 (8.2%)	1053	113 (10.7%)	RR 0.77 (0.54 to 1.09)	p=0.14	42% (p=0.19)
Hypoglycaemic events  (SGLT2i studies)	1092	6 (0.5%)	1053	8 (0.8%)	RR 0.76 (0.26 to 2.24)	p=0.62	0% (p=0.38)

^*p<0.05.

^†mean.

BMI, body mass index; GDMT, guideline-directed medical therapy; LVEF, left ventricular ejection fraction; MD, mean difference; NYHA, New York Heart Association; RR, relative risk; SGLT2i, sodium-glucose cotransporter-2 inhibitor.

More than 1/3 of older HFrEF patients were NYHA class III/IV (pooled prevalence 33.8%, 95% CI 22.4 to 46.2, I² 100%) (table 2, online supplemental appendix 9). Older patients who received GDMT were less likely to be in NYHA class III/IV HF (RR 0.8, 95% CI 0.75 to 0.94, I² 80%, Tau² 0.02, p=0.002) (table 3, online supplemental appendix 10). Left ventricular ejection fraction was similar between the GDMT and no GDMT groups (mean difference −0.60 (-1.96–0.77), I² 92%, Tau² 4.55, p=0.39) (table 3, online supplemental appendix 10).

Study-level data on the use of each GDMT components (including ACEi/ARB/ARNI, BB, MRA and SGLT2i) were presented in online supplemental appendix 8. The pooled prevalence of ACEi/ARB/ARNi among older patients with HFrEF was 74.2% (95% CI 66.6 to 81.1, I² 100%). The pooled prevalence of BB was 65.1% (95% CI 50.3 to 78.6, I² 100%), MRA pooled prevalence was reported as 33.6% (95% CI 23.5 to 44.5, I² 100%) and SGLT2i 16.4% (95% CI 0.0 to 41.6, I² 100%) (table 2, online supplemental appendix 11).

Adverse events

Four studies^{28 30 31 35} reported adverse drug events. Older patients receiving BB were more likely to develop bradycardia (two studies:^{28 31} RR 3.98 (2.72–5.82), I² 0%, p<0.001) (table 3, online supplemental appendix 10). There was no difference in the development of hypotension, acute renal impairment and hypoglycaemia between the GDMT and placebo group (hypotension (RR 2.98 (0.21–43.23), I² 73%, p=0.42); acute renal impairment (RR 0.77 (0.54–1.09), I² 42%, p=0.14); hypoglycaemic events (RR 0.76 (0.26–2.24), I² 0%, p=0.62)) (table 3, online supplemental appendix 10).

OS

Pooled analysis demonstrated a survival benefit of GDMT in older patients with HFrEF (34 studies,^24–57 HR 0.69 (0.64–0.74) p<0.001, Tau² 0.03, I² 71%) although statistical heterogeneity was high (I² 71%, between-study variance was low (Tau² 0.03) indicating that while the proportion of variability attributed to heterogeneity was high, the absolute differences in treatment effect across studies were small. Therefore, despite the high I², the pooled HR can be interpreted with reasonable confidence (figure 2). Subgroup analysis according to age cut-off, number of GDMT classes used, year of evolution of HF treatment, study type, LVEF cut-off, risk of bias, sample size and geographical location does not explain heterogeneity (figure 3, online supplemental appendix 12). The exclusion of three observational studies^{25 40 50} of high RoB did not change the pooled effect estimates: similarly, subgroup analysis stratified by RoB yielded consistent findings (online supplemental appendix 13), suggesting that overall primary outcome was not driven by methodologically weaker studies.

Figure 2. Forest plot to compare the overall survival between the GDMT versus the no GDMT group in the older HFrEF population. GDMT, guideline-directed medical therapy; HFrEF, heart failure with reduced ejection fraction.

Figure 3. Summary of subgroup analyses according to (1) age cut-off, (2) number of GDMT, (3) year of evolution of heart failure treatment, (4) study type, (5) LVEF cut-off, (6) risk of bias, (7) sample size and (8) geographical locations. GDMT, guideline-directed medical therapy; LVEF, left ventricular ejection fraction.

Meta-regression demonstrated no evidence of moderation by age (β=0.0061; p=0.45) or by the number of GDMT agents (two agents (β=0.014; p=0.14); three agents (β=0.0061; p=0.88)). These findings indicate that neither age nor GDMT intensity accounted for between-study differences in effect estimates (online supplemental appendix 14). Begg’s rank correlation tests suggested evidence of small-study effects (τ = –0.32, p=0.008). However, Egger’s regression intercept did not reach significance (intercept = –0.26, p=0.076). The funnel plot showed some scatter among smaller studies, but no overt asymmetry on visual inspection (online supplemental appendix 15). Trim-and-fill imputed eight potentially missing studies, and while the adjusted pooled effect was attenuated compared with the primary model, it remained in the same direction (online supplemental appendix 16). Taken together, these findings indicate that small-study effects are likely present. Although the adjusted effect remained directionally similar, these findings indicate possible publication bias that may have inflated the pooled effect size. Overall, treatment with GDMT is associated with improved survival compared with no GDMT among older patients with left ventricular systolic dysfunction. There was evidence of small study effects based on funnel-plot asymmetry and trim-and-fill analyses. Given this indication of possible publication bias, the certainty of evidence, using the GRADE framework, was downgraded for reporting bias (online supplemental appendix 17).

Discussion

This meta-analysis included 34 studies with 92 916 patients over the age of 65 and found a survival benefit of GDMT in older patients with HFrEF.

HFrEF remains a major healthcare burden and is expected to rise further, driven largely by population growth and ageing.⁴ Among the four pillars of GDMT for HFrEF, this study suggests that the most frequently prescribed class in older adults was one of ACEi/ARB/ARNI at over 70%. BB use was also relatively high at approximately 65%, whereas MRA uptake was just over 30%, in part due to the perceived fear of hyperkalaemia, particularly in patients with chronic kidney disease at baseline.⁵⁸ Not unexpectedly, SGLT2i use was low, at just above 15%, likely reflecting their more recent incorporation into GDMT in the 2021 ESC and the 2022 AHA guidelines. Nonetheless, both DAPA-HF³⁵ and EMPEROR-Reduced³⁰ trials conducted subgroup analyses looking at the treatment effects in the older group with an age cut-off of 65 years and reported beneficial effects.

While GDMT confers a survival benefit, older patients remain under-represented in well-conducted clinical trials.⁵⁹ Large registries show that older HFrEF patients are less frequently prescribed GDMT than younger individuals and timely optimisation remains suboptimal in real-world practice.⁷ Barriers include physiological limitations,⁶⁰ lower drug tolerability⁶¹ and concerns regarding adverse drug reactions.⁶² Achieving guideline-recommended target doses is particularly challenging in those aged ≥80 years, where the relevance of aggressive up-titration is uncertain given the increased risk of hypotension, bradycardia, falls and related complications. Consistent with prior pivotal HFrEF trials,⁶³ this review also noted a higher incidence of bradycardia with BB in older adults. These challenges underscore the need for structured implementation strategies, including multidisciplinary HF teams and algorithm-driven titration with clinician audit and feedback. Notably, a meta-analysis of 17 landmark HFrEF trials found similar adverse-events rates between GDMT and placebo, suggesting many events may reflect the intrinsic high-risk nature of HF phenotype rather than drug-specific effects.⁶⁴ Nonetheless, high-quality studies focused on older adults are needed to better define the benefit-risk profile of GDMT, as side effects remain a major driver of treatment discontinuation and the safety in older patients remains uncertain.

GDMT is frequently under-used in frail individuals with HF. While some studies^{25 33} reported that frailty status did not interact with the adverse effects of GDMT, others have demonstrated an association between frailty and substantially higher risk for adverse outcomes.⁶⁵ The prevalence of frailty among patients with HF varies from 19% to 52%.⁶⁶ Importantly, individuals with HF are up to six times more likely to be frail, and frail individuals have a significantly higher risk of developing HF with a bidirectional relationship between the two conditions⁶⁷ in which HF-specific mechanisms may contribute to frailty development and vice versa. Hamada et al⁶⁸ further showed that frailty was associated with lower use of standard GDMT, even after adjusting for hypotension, renal impairment, bradycardia and hyperkalaemia. Similarly, our study found that older patients with HFrEF and a higher burden of frailty were less likely to be initiated on GDMT. However, the proportion of frail individuals included across the contributing studies of this meta-analysis was low at only 7%. This misalignment between the clinical population of greatest interest and the available evidence indicates that conclusions regarding safety in older frail patients are underpowered, thereby limiting the strength of any definitive conclusion.

Therein lies the importance of close monitoring during GDMT initiation and titration in older adults to allow early detection of adverse events and timely implementation of strategies to improve tolerability, particularly in the frail population.⁶⁹ Specific approaches to modify elements of frailty phenotype should be considered, including (1) cardiac rehabilitation⁷⁰ to address low physical activity, weakness and exhaustion and (2) nutritional optimisation⁷¹ to target metabolic dysregulation, sarcopenia and unintentional weight loss. Therapeutic goals in frail older patients may differ from those established in younger HFrEF populations, and frailty-directed interventions may have important prognostic implications.⁷²

Limitations

First, funnel plot asymmetry and trim-and-fill imputation of eight potentially missing studies indicates the presence of small study effects suggesting that publication bias cannot be excluded. Second, this review included secondary data analysis of nine RCTs. Given that data were originally collected for a different purpose, this introduces the risk of selection bias. A third limitation was none of the included studies evaluated all four foundational pillars of GDMT (ARNI/ACEi/ARB, BB, MRA and SGLT2i) concurrently. This absence of comprehensive GDTM implementation limits extrapolation to current standards of care. A fourth was the limited number of studies that reported on whether target dose of each GDMT class was achieved in the GDMT group. A fifth limitation includes the inherent assumptions of meta-analysis—we were limited to data reported in published articles and had limited access to individual patient level data which would have been particularly helpful for our subgroup analyses and the assessment of recurrent hospitalisation and quality of life which are important in the HFrEF population. A fifth limitation is that we are unable to isolate the effect of individual drug class because several studies administered combination GDMT regimens (two or more classes) versus no GDMT, which prevents reliable isolation of individual class effect. Due to the low response rate from contacted authors to retrieve additional data for secondary outcomes, this raises the possibility of reporting bias, and the exclusion of studies with incomplete reporting may have reduced statistical power. The limitations of this study emphasise that conclusions regarding GDMT in older HFrEF patients cannot be made definitively.

Conclusion

The use of GDMT in older patients with LV systolic dysfunction was associated with a survival benefit. However, frailty may impede the successful implementation of these therapies. High-quality studies are urgently needed to develop targeted strategies and optimise the management of older, frail patients with LV systolic function.

Data availability statement

Data are available in a public, open access repository. All data relevant to the study are included in the article or uploaded as supplementary information.