Efficacy and safety of stem cell therapy for myocardial infarction and heart failure: an updated systematic review and meta-analysis of randomized controlled trials

Abstract

Background/objectives

Cardiovascular diseases (CVDs), particularly myocardial infarction (MI) and heart failure (HF), remain global causes of morbidity and mortality, despite the current treatment options. Stem cell therapy (SCT) has emerged as a promising intervention aimed at halting disease progression and promoting cardiac repair. Nonetheless, the clinical efficacy, optimal cell type, delivery method, and safety profile of SCT remain inadequately defined.

Methods

This systematic review and meta-analysis aimed to evaluate the efficacy and safety of SCT in patients diagnosed with ischemic heart disease (IHD) and HF. PubMed, Web of Science, Embase, Scopus, and Science Direct were searched to retrieve randomized controlled trials (RCTs) investigating SCT in patients with MI or HF. Primary outcomes encompassed changes in left ventricular ejection fraction (LVEF), end-diastolic and end-systolic volumes (LVEDV, LVESV), infarct size, functional status, and quality of life measures. Risk ratios were calculated for safety outcomes. Subgroup analyses were executed based on follow-up duration, delivery method, and type of stem cell utilized.

Results

This review included 35 RCTs comprising 3345 patients (1875 in the SCT group and 1488 in the control group). The SCT indicated that significantly enhanced LVEF at 3, 6, and 12 months (mean difference [MD] = 1.43; 95% CI 0.92 to 1.95; p < 0.00001), while simultaneously reducing LVEDV (MD = − 5.23; 95% CI − 7.55 to − 2.91; p < 0.0001) and LVESV (MD = − 6.91; 95% CI − 9.01 to − 4.82; p < 0.00001). Additionally, infarct size demonstrated significant reductions at 6 and 12 months. Patients undergoing SCT exhibited improvements in functional status and quality of life. The safety profile of SCT indicated that it was well tolerated.

Conclusions

SCT appears to be a safe and modestly effective adjunctive therapy for patients with IHD and HF. The standardization of treatment protocols and the conduct of longer-term studies are critical to validate its clinical utility and optimize therapeutic outcomes.

Systematic review registration: PROSPERO CRD42024582716

Supplementary Information

The online version contains supplementary material available at 10.1186/s13643-026-03133-w.

Introduction

CVDs are a group of conditions that affect the cardiovascular system and its supporting structures. The most common CVDs include coronary artery disease (CAD), congestive heart failure (CHF), ischemic and non-ischemic cardiomyopathy, as well as peripheral arterial disease (PAD), among others [1]. IHD remains a significant global health concern, accounting for 17.9 million deaths annually and contributing to nearly two-thirds of systolic heart failure cases [2, 3]. Although current pharmacological and interventional treatments help manage symptoms and limit adverse cardiac remodeling, they remain largely ineffective in reversing the underlying issue of irreversible heart tissue loss [4, 5].

MI is the leading cause of mortality globally despite substantial advancements in treatment, ventricular dysfunction continues to be the primary contributor to morbidity and mortality in these patients [6]. Approximately 14–36% of patients hospitalized for acute myocardial infarction (AMI) develop HF [7]. Additionally, post-MI heart failure continues to be associated with high morbidity and mortality rates [5], with current treatments failing to regenerate damaged cardiac tissue, highlighting the need for innovative therapeutic approaches [8]. Significant myocardial remodeling and chamber dilation are associated with worse prognosis, and conventional treatments may not be effective for these patients, in such cases, cardiac transplantation or long-term mechanical circulatory support may be considered; however, they come with significant risks and are further restricted by limited donor availability, patient eligibility criteria, and high costs [9, 10].

Cell therapy has emerged as a potential adjunctive approach to improve cardiac function and alleviate heart failure symptoms. Several small-scale RCTs have tested different cell populations in patients with ischemic heart failure and refractory angina. While all have confirmed their safety, clinical efficacy findings remain inconsistent [11–15]. This approach has been explored as a potential method to partially repair myocardial injury and improve cardiac function following a MI [16, 17]. Intracoronary administration of various cell populations, including circulating progenitor cells, bone marrow-derived progenitor cells, bone marrow cells, peripheral blood stem cells, hematopoietic stem cells, and allogeneic bone marrow mesenchymal stromal cells, has been explored in the treatment of AMI, with some studies reporting promising outcomes [18–20]. While preclinical and certain clinical studies suggest potential benefits in improving left ventricular (LV) function, myocardial perfusion, and reducing infarct size, the overall effectiveness of cell therapy remains controversial, as many clinical trials have failed to demonstrate significant improvements in left ventricular ejection fraction (LVEF) [16, 17, 21–23]. This meta-analysis aims to examine the effectiveness, safety, and clinical impact of stem cell therapy in managing MI and HF.

Methods

This review was conducted and reported following the guidelines set by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [24]. The study protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) (ID: CRD42024582716).

Search strategy

Two independent researchers searched PubMed, Embase, Web of Science, Scopus, and Science Direct databases from inception until August 2024. The search strategy incorporated relevant keywords and Medical Subject Headings (MeSH) related to heart failure, myocardial infarction, and stem cells. Further details of the search strategy, including the specific MeSH terms and database-specific search strings used for each search engine, are provided in the Appendix.

Study eligibility criteria

This review established the following eligibility criteria for inclusion: (1) participants must be aged 18 or older with a diagnosis of IHD or CHF; (2) interventions must involve stem cell-based therapies, regardless of their type, dosage, or administration method; (3) control groups should consist of conventional care, alternative treatments for IHD, or placebo comparators; (4) studies must assess efficacy and/or safety outcomes, which should include metrics such as LVEF, LVEDV, LVESV, infarction size, N-terminal pro-B-type natriuretic peptide (NT-proBNP)levels, quality of life assessments, all-cause mortality, overall adverse events (AEs), serious adverse events (SAEs), total hospitalization rates, cardiovascular-related hospitalization rates, and cardiac transplantation rates; (5) eligible studies should have been published from inception up to August 2024; and (6) only RCTs published in English were considered.

Exclusion criteria were defined as follows:

(1) Any interventions that are not stem cell-based; (2) studies focusing on conditions unrelated to ischemic heart disease; (3) study designs that do not conform to the parameters of English-written RCTs; (4) outcome measures that do not align with the specified assessment criteria.

Study selection

All studies identified via electronic database searches were imported into Rayyan Software 5.2 for initial screening and duplicate removal [25]. The titles and abstracts underwent independent assessment by six authors based on predetermined inclusion criteria. Discrepancies during this phase were resolved through discussions with a seventh reviewer. Subsequently, the full texts of all potentially eligible studies were retrieved and further evaluated for eligibility by two reviewers independently. Reasons for exclusion during the full-text review phase were documented. The entire study selection process was summarized using a PRISMA flow diagram.

Data extraction

Five reviewers conducted independent data extraction from the selected studies using a standardized data extraction form. Discrepancies in the data were addressed and resolved through discussions with the sixth reviewer. The extracted data included various study characteristics such as author details, country of origin, study design, sample size, publication year, and inclusion and exclusion criteria. Patient demographic information captured encompassed age and sex. Furthermore, data pertinent to stem cell interventions was meticulously recorded, detailing the type, dosage, treatment regimen, route of administration, adjunct therapies, total dosage, and discharge medications. Additionally, the follow-up periods for the studies were documented.

Data synthesis and analysis

The statistical analysis was performed using Review Manager (RevMan) software, Version 5.4.1, developed by The Nordic Cochrane Centre, part of The Cochrane Collaboration, in 2020 [26]. Mean differences (MDs) were calculated for the mean change at baseline, 3, 6, and 12 months intervals for the LVEF, LVEDV, LVESV, 6-min walk test (6MWT), Infarction Size, New York Heart Association (NYHA), N-terminal pro-B-type natriuretic peptide (NT-proBNP), Kansas City Cardiomyopathy Questionnaire (KCCQ), Minnesota Living With Heart Failure Questionnaire (MLHFQ), and Global Longitudinal Strain (GLS). Furthermore, for dichotomous outcomes, we calculated risk ratios (RR) along with 95% confidence intervals (CI) for safety outcomes including all-cause mortality, total hospitalization, cardiovascular hospitalization, total adverse events, heart transplants or cardiac devices, MI, serious adverse events, ventricular arrhythmia, and worsening heart failure. The findings from studies were displayed using forest plots, and a funnel plot was generated to evaluate publication bias across all primary outcomes. To assess the levels of heterogeneity, we employed Higgin's I² statistic. Despite substantial heterogeneity (I² > 70%) in some analyses, pooled estimates were reported using a random-effects model, acknowledging that the variability likely reflects both methodological and clinical differences across studies.

Heterogeneity exploration

To examine the impact of clinical heterogeneity on the outcomes of the meta-analysis, we conducted subgroup analyses stratified by time intervals, routes of administration, and types of stem cells used.

Results

Search results

A flowchart illustrating the study selection process is presented in (Fig. 1). A total of 1359 records were initially identified through electronic database searches. After removing 172 duplicate records, 1187 unique records remained for title and abstract screening. Of these, 1120 were excluded due to the absence of the outcome of interest. The full text of the remaining 67 articles was assessed for eligibility. Following this assessment, 24 studies were excluded for being irrelevant to stem cell–based therapy in ischemic heart disease, and 8 studies were excluded due to unavailable full-text articles. Ultimately, 35 studies were included in the qualitative and quantitative analysis.

Fig. 1.

PRISMA flowchart of the identification of the studies

Study characteristics

Appendix 1 Tables A1 and A2 present the baseline characteristics of studies investigating the use of SCT in ischemic heart disease and heart failure across various international populations. The review includes 35 RCTs conducted between 2004 and 2024 in countries including Spain, Slovenia, Iran, Denmark, the USA, China, South Korea, and others.

Sample sizes varied widely across studies. The number of patients receiving SCT ranged from 5 participants [27] to 283 participants [28], while control groups ranged from 4 [27] to 282 patients [28]. Sex distribution differed between studies; for example, the treatment group in Attar et al. (2023) [29] was 100% male, whereas Perin et al. (2011) [30] reported the highest percentage of females in the treatment group, which was 50%.

Baseline clinical measures, including estimated glomerular filtration rate (eGFR), low-density lipoprotein cholesterol (LDL-C), and cardiovascular medication use, were variably reported. For example, baseline LDL-C values ranged from 1.2 mmol/L [31] to 3.1 mmol/L [32]. Many studies reported the use of key cardiovascular medications, including angiotensin-converting enzyme (ACE) inhibitors, beta-blockers (β-blockers), mineralocorticoid receptor antagonists (MRAs), statins, and sodium-glucose cotransporter-2 (SGLT2) inhibitors. Statin use was as high as 97.5% in the intervention group of Perin et al. (2023) [30], whereas in some trials (e.g., He et al., 2020) [33], specific drug use was limited or unreported.

Overall, the studies included a heterogeneous patient population with diverse demographic and clinical characteristics, highlighting the diverse contexts in which SCT has been explored.

Risk of bias assessment

We assessed the quality of the included studies using the Cochrane risk of bias (RoB 2.0) tool [34]. The results indicated that most studies had a low risk of bias in the domains of outcome measurement and missing data. The randomization process and adherence to intended interventions were generally low risk, although some studies raised concerns or showed high risk. A considerable proportion of studies exhibited high risk and moderate concerns in the selection of reported results. Regarding overall bias, the majority of studies were deemed to be at risk, with only a few rated as low risk. A summary and visual representation of the risk of bias are presented in Fig. 2 below.

Fig. 2.

Risk of bias graph (A) and summary (B) of the quality assessment of the included Studies [34]

Publication bias

Figure 3 shows funnel plots for the primary outcomes including the comparison of mean change in the LVEF, LVEDV, and LVESV. For LVEDV, and LVESV were symmetric and suggested no obvious publication bias. In contrast, LVEF revealed possible publication bias at 3, and 6 months.

Fig. 3.

Funnel plot of studies assessing publication bias. A Ejection fraction. B Left Ventricle End-Diastolic Volume Index. C Left Ventricular End-Diastolic Volume Index

Evaluation of efficacy outcomes at 3, 6, and 12 months

Echocardiographic parameters (LVEF, LVEDV, LVESV) were considered surrogate markers of cardiac remodeling and function.

LVEF

Figure 4 presents the findings from our meta-analysis assessing the impact of SCT on LVEF at multiple time intervals. This analysis confirms that the baseline LVEF was statistically equivalent between the SCT and placebo groups, thereby validating the effectiveness of the randomization process in ensuring balance between the study cohorts. The 3-month endpoint revealed (MD = 1.29; 95% CI 0.44 to 2.15; p = 0.003), indicating that SCT significantly enhances LVEF when compared to the placebo. At the 6-month evaluation, a mean difference of 2.23 (95% CI 1.19, 3.27, p < 0.00001) was reported. Furthermore, at the 12-month follow-up, a mean difference of 2.31 (95% CI 1.21, 3.41, p < 0.00001) was indicated, suggesting that the improvement in LVEF is sustained and increases over time. The overall effect yielded a pooled mean difference of 1.43 (95% CI 0.92, 1.95, p < 0.00001) favoring SCT.

Fig. 4.

A pooled analysis comparing the effects of SCT vs. placebo on LVEF at baseline, 3 months, 6 months, and 12 months

LVEDV

Figure 5 shows that the analysis of LVEDV at baseline reveals statistically equivalent results between the SCT and placebo groups. Notably, SCT was associated with a consistent and statistically significant reduction in LVEDV at the 3-month interval (MD = − 3.20; 95% CI − 5.80, − 0.60, p = 0.02), at the 6-month interval (MD = − 8.81; 95% CI − 12.55, − 5.07, p < 0.00001), and the 12-month mark (MD = − 8.16; 95% CI − 12.66, − 3.65, p = 0.0004). The overall effect demonstrated a mean difference of − 5.23 (95% CI − 7.55, − 2.91, p < 0.0001), indicating a beneficial impact of SCT on cardiac remodeling.

Fig. 5.

A pooled analysis comparing the effects of SCT vs. placebo on LVEDVI at baseline, 3 months, 6 months, and 12 months

LVESV

Figure 6 illustrates the analysis of LVESV, which demonstrates that the administration of SCT results in a significant reduction in LVESV across all time-based subgroups. The overall pooled effect indicates a mean difference of − 6.91 (95% CI − 9.01, − 4.82, p < 0.00001), favoring SCT.

Fig. 6.

A pooled analysis comparing the effects of SCT vs placebo on LVESVI at baseline, 3 months, 6 months, and 12 months

Functional capacity (6MWT)

Figure 7 shows an analysis of the 6MWT, which was not significant across all time-based subgroups.

Fig. 7.

A pooled analysis comparing the effects of SCT vs. placebo on 6MWT at baseline, 3 months, 6 months, and 12 months

Infarction size

Figure 8 illustrates that infarct size remained comparable at baseline and the 3-month mark. However, notable reductions were documented at 6 months (MD = − 1.80; 95% CI − 2.76 to − 0.84; p = 0.0002) and 12 months (MD = − 2.56; 95% CI − 4.47 to − 0.65; p = 0.009). The overall effect favored SCT (MD = − 1.57; 95% CI − 2.13 to − 1.01; p < 0.00001).

Fig. 8.

Forest plot for the pooled analysis of the difference in the mean change in infarction size at baseline, 3-month, 6-month, and 12-month intervals

NYHA functional class

Figure 9 shows that the NYHA class scores at baseline, 3 months, and 12 months were similar; however, SCT demonstrated a significant enhancement at 6 months (MD = − 0.58; 95% CI − 1.09 to − 0.06; p = 0.03). The overall effect was statistically significant (MD = − 0.36; 95% CI − 0.53 to − 0.19; p < 0.0001).

Fig. 9.

Forest plot for the pooled analysis of the difference in the mean change in NYHA class at baseline, 3-month, 6-month, and 12-month intervals

NT-proBNP

No significant differences were detected between groups across all time points, as depicted in Fig. 10.

Fig. 10.

Forest plot for the pooled analysis of the difference in the mean change in NT-proBNP at baseline, 3-month, 6-month, and 12-month intervals

Quality of life (KCCQ and MLHFQ)

Figure 11 shows an analysis of the KCCQ, revealing that SCT administration resulted in significant improvements in KCCQ scores across all time-based subgroups; however, the overall pooled effect did not show statistical significance (MD = 3.72; 95% CI − 0.14, 7.58, p = 0.06). In contrast, MLHFQ scores were significantly improved in the SCT group overall (MD = − 4.60; 95% CI − 8.25 to − 0.94; p = 0.01; Fig. 12), despite missing data at 3 months.

Fig. 11.

Forest plot for the pooled analysis of the difference in the mean change in KCCQ at baseline, 3-month, 6-month, and 12-month intervals

Fig. 12.

Forest plot for the pooled analysis of the difference in the mean change in MLHFQ at baseline, 3-month, 6-month, and 12-month intervals

GLS

No significant differences in GLS were observed between groups at any follow-up (Fig. 13).

Fig. 13.

Forest plot for the pooled analysis of the difference in the mean change in GLS at baseline, 3-month, 6-month, and 12-month intervals

Effectiveness by stem cell type

This meta-analysis evaluates the efficacy of various stem cell therapies in comparison to placebo regarding several cardiac functional parameters

LVEF

As shown in Fig. 14, CD34⁺ cells were the only type with significant benefit at 3 months (MD = 3.66; 95% CI 0.08 to 7.24; p = 0.04). At 6 months, bone marrow-derived mesenchymal stem/stromal cells (MSCs) (MD = 2.20; 95% CI 0.37 to 4.02; p = 0.02), mesenchymal precursor cells (MD = 2.85; 95% CI 0.18 to 5.52; p = 0.04), and CD34⁺ cells (MD = 4.24; 95% CI 2.57 to 5.91; p < 0.00001) demonstrated efficacy. At 12 months, 4 out of 10 stem cell types demonstrated a significant improvement in ejection fraction compared to placebo, including c-kit+ cardiac cells (MD = 1.20; 95% CI 0.75, 1.65, p < 0.00001), bone marrow-derived MSCs (MD `= 2.20; 95% CI 0.37, 4.02, p = 0.02), mesenchymal precursor cells (MD = 2.58; 95% CI 0.18, 5.52, p = 0.04), and CD34⁺ cells (MD = 4.24; 95% CI 2.57, 5.91, p < 0.00001).

Fig. 14.

Forest plot summarizing the effects of different stem cell types on left ventricular remodeling in patients with ischemic heart disease. Ejection Fraction (EF%) A At 3 months. B At 6 months. C At 12 months

LVEDVI

As shown in Fig. 15, significant reductions were seen at 3 months with mesenchymal precursor cells exhibited a statistically significant difference from the placebo group, with a mean difference of − 6.56 (95% CI − 10.60 to − 2.52, p = 0.001). At the 6-month follow-up, significant improvements compared to the placebo group were observed in adipose-derived MSCs (MD = − 7.11; 95% CI − 13.86, − 0.36, p = 0.04), cardiopoietic MSCs (MD = − 6.68; 95% CI − 11.70, − 1.65, p = 0.009), and c-kit+ cardiac cells (MD = − 4.30; 95% CI − 5.52, − 3.08, p < 0.00001). At the 12-month time point, four out of eight stem cell types demonstrated significant improvement, including c-kit+ cardiac cells (MD = − 6.60; 95% CI − 8.14, − 5.06, p < 0.00001), umbilical cord-derived MSCs (MD = − 4.67; 95% CI − 6.85, − 2.49, p < 0.0001), JV5-100 (MD = − 16.40; 95% CI: − 21.34, − 11.46, p < 0.0001), and cardiopoietic MSCs (MD = − 14.60; 95% CI − 21.29, − 7.91, p < 0.0001).

Fig. 15.

Forest plot summarizing the effects of different stem cell types on left ventricular remodeling in patients with ischemic heart disease. Left Ventricular End-Diastolic Volume Index (LVEDVI, mL/m2) A At 3 months. B At 6 months. C At 12 months

LVESVI

Figure 16 shows that three out of six types of stem cells show significant improvement at 3-month intervals mesenchymal precursor cells (MD = − 8.43;95% CI − 11.56, − 5.30, p < 0.00001). Bone marrow-derived MSCs (MD = − 5.90; 95% CI − 10.77, − 1.031, p = 0.01), and cardiopoietic MSCs (MD = -−.00; 95% CI − 10.23, 0.23, p = 0.06). At 6 months just two out of seven stem cells showed a significant difference c-kit+ cardiac cells (MD = − 6.36;95% CI − 7.53, − 5.19, p < 0.00001), and mesenchymal precursor cells (MD = − 6.92;95% CI − 9.81, − 4.04, p < 0.00001). Lastly, at a 12-month interval, six out of nine stem cells demonstrate a significant difference (MD = − 9.65; 95% CI − 10.97, − 8.33, p < 0.00001), JVS-100 (MD = − 29.30; 95% CI − 33.56, − 25.04, p < 0.00001), umbilical cord-derived MSCs (MD = − 4.55; 95% CI − 5.23, − 3.87, p < 0.00001), mesenchymal precursor cells (MD = − 6.85; 95% CI − 9.14, − 4.56, p < 0.00001), bone marrow mononuclear cells/autologous bone marrow mononuclear cells (BMMNCS/ABMMNCS) (MD = − 4.55; 95% CI − 29.22, − 13.78, p < 0.00001), and cardiopoietic MSCs (MD = − 11.00; 95% CI − 16.77, − 5.23, p = 0.0002). The overall effect indicates that, across the outcomes and all time intervals, SCT demonstrates a statistically significant improvement compared to the placebo.

Fig. 16.

Forest plot summarizing the effects of different stem cell types on left ventricular remodeling in patients with ischemic heart disease. Left Ventricular End-Systolic Volume Index (LVESVI, mL/m²) A At 3 months. B At 6 months. C At 12 months

Effectiveness by route of administration

Our meta-analysis assesses the efficacy of various routes of administration for SCT in comparison to placebo, focusing on several cardiac functional parameters.

LVEF

Figure 17A indicates no significant differences among delivery methods at 3 months. At 6 months, intracoronary infusion (MD = 4.51; 95% CI 1.78, 7.24, p = 0.001), transendocardial injection (MD = 3.16; 95% CI 1.07, 5.26, p = 0.003), and surgical implantation (MD = 4.90; 95% CI 1.58, 8.22, p = 0.004) demonstrated significant differences favoring SCT. Furthermore, at the 12-month follow-up, intracoronary infusion (MD = 2.51; 95% CI 0.59, 4.43, p = 0.01), transendocardial injection (MD = 1.46; 95% CI 0.52, 2.40, p = 0.002), and intramyocardial injection (MD = 3.27; 95% CI 0.49, 6.05, p = 0.02) also exhibited significant differences.

Fig. 17.

Forest plot summarizing the effects of different stem cell types on left ventricular remodeling in patients with ischemic heart disease. A Ejection fraction (EF%). B Left Ventricular End-Diastolic Volume Index (LVEDVI, mL/m²). C Left Ventricular End-Systolic Volume Index (LVESVI, mL/m²) at 3, 6, and 12 months

LVEDV

As shown in Fig. 17B, significant LVEDVI reductions were observed with intravenous infusion (MD = − 6.56; 95% CI − 10.60, − 2.52, p = 0.01). Additionally, at the 6-month interval, both intramyocardial injection (MD = − 3.47; 95% CI − 5.50, − 1.44, p=0.0008) and subcutaneous mobilization (MD = − 6.00; 95% CI − 10.40, − 1.60, p = 0.008), along with surgical implantation (MD = − 29.50; 95% CI − 40.02, − 18.98, p < 0.00001), exhibited significant differences. Furthermore, at the 12-month follow-up, intramyocardial injection (MD = − 7.41; 95% CI − 13.60, − 1.21, p = 0.02) and intravenous infusion (MD = − 7.40; 95% CI: −12.34, − 2.46, p = 0.003) also demonstrated significant differences.

LVESV

Figure 17C reveals significant reductions with intravenous infusion (MD = − 8.43; 95% CI − 11.56, − 5.30, p < 0.00001). At the 6-month mark, intravenous infusion (MD = − 7.40; 95% CI − 11.19, − 3.61, p = 0.0001), surgical implantation (MD = − 2.98; 95% CI − 4.94, − 1.02, p = 0.003), transendocardial injection (MD = − 8.65; 95% CI − 14.86, − 2.45, p = 0.006), intramyocardial injection (MD = − 12.09; 95% CI − 22.67, − 1.51, p = 0.03), and subcutaneous mobilization (MD = − 9.00; 95% CI − 12.40, − 5.60, p < 0.00001) showed statistically significant differences favoring SCT. These trends continued at 12 months for both intramyocardial and intravenous infusion routes.

$$A-EF\%B-LVEDVI(mL/M^2)C-LVESVI(mL/m{.}^2)$$

Safety outcomes

Figure 18 and Table 1 present a comparison of intervention and control groups concerning various adverse events. A total of 24 studies reported safety outcomes. The results indicate that the intervention groups exhibited a lower incidence of all-cause mortality (5.56% compared to 6.76%), total hospitalization (22.31% versus 28.89%), cardiovascular hospitalization (18.79% compared to 25.54%), total adverse events (43.36% versus 50.73%), the need for heart transplants or cardiac devices (12.30% versus 18.30%), and myocardial infarction (4.54% compared to 4.68%). However, it is noteworthy that certain adverse events were more prevalent in the intervention group, including serious adverse events (48.58% versus 47.06%), ventricular arrhythmia (13.40% versus 10.86%), and worsening heart failure (10.76% compared to 10.30%).

Fig. 18.

Forest plot summarizing the safety outcomes of SCT. A All-cause mortality. B Total hospitalization. C Cardiovascular hospitalization. D Total adverse events. E Serious adverse events. F Ventricular arrhythmia. G Heart transplant or cardiac device. H Heart failure worsening. I Myocardial infarction

Table 1.

Summary of key findings of safety outcomes

Discussion

Efficacy of stem cell therapy: functional and structural benefits

This systematic review and meta-analysis demonstrates that, across 35 randomized controlled trials, SCT is associated with statistically significant improvements in LVEF, with a pooled mean increase of 1.43%, accompanied by modest reductions in LVEDV and LVESV, most consistently observed at 6 and 12 months.

However, the clinical significance of these echocardiographic changes remains uncertain. There is no universally accepted minimal clinically important difference (MCID) for LVEF in heart failure populations, and a pooled increase of 1.43% is unlikely, in isolation, to translate into meaningful changes in patient management or prognosis. Current heart failure guidelines therefore emphasize assessment of LVEF trajectory over time rather than reliance on single time-point measurements [35].

Several trials also reported improvements in functional and surrogate outcomes, including NYHA class, exercise capacity, and biomarkers such as NT-proBNP. For example,

Qayyum et al. [31] documented reductions in NT-proBNP and improvements in NYHA class following intramyocardial injection of adipose-derived mesenchymal stem cells, while Perin et al. [28] reported improvements in 6-min walk test performance and reduced heart failure–related hospitalizations with mesenchymal precursor cells in a large trial (n = 565).

Patient-reported outcomes offer a more direct measure of clinical relevance, though they are inherently subjective. MCIDs are better established for measures such as the Kansas City Cardiomyopathy Questionnaire (KCCQ), where changes of approximately 5–10 points are considered meaningful. Although some trials reported improvements in quality of life, these outcomes were inconsistently measured and often underpowered.

Overall, SCT appears to confer modest yet reproducible structural and functional benefits. The clinical impact at the population level, however, is likely limited, underscoring the need for future trials powered to detect meaningful patient-centered outcomes alongside echocardiographic changes over longer follow-up periods.

Safety profile

Stem cell therapy appears to be relatively well tolerated, though caution is warranted in interpreting safety data. Most trials reported reassuring safety profiles, with serious complications being rare and, where present, numerically lower in the SCT group compared with controls. Minor adverse events, such as local inflammation or transient fever, were reported in some trials, including He et al. (2020) [33], Sung et al. (2020) [36], and Kim et al. (2018) [37]. Notably, no evidence of tumorigenicity or immune-mediated rejection has been observed, supporting the immune-privileged status of many stem cell therapies, particularly mesenchymal stem cells.

Nonetheless, certain studies raised potential safety concerns, particularly regarding ventricular arrhythmias, worsening heart failure, and procedure-related complications. These observations are biologically plausible, as intramyocardial or transendocardial delivery may transiently disrupt local myocardial tissue or electrical conduction, and stem cells themselves can provoke short-lived inflammatory responses.

Although overall adverse events did not reach statistical significance, some studies reported numerically higher rates of serious events. For example, Perin et al. (2023) observed lower cardiovascular hospitalization in the SCT group (170 vs. 191), yet the number of patients experiencing worsening heart failure and myocardial infarction was notably higher in the intervention arm (7 vs. 2 for both outcomes).

These findings emphasize that SCT may not be universally safe for all cardiac patients and highlight the need for careful interpretation of results. In addition, they reinforce the importance of careful patient selection, close interval monitoring following procedures, development of a standardized review process for serious adverse events, and extended patient follow-up in future studies.

Type and source of stem cells

The cell types used across the included studies were diverse, with mesenchymal stem/stromal cells (MSCs) being the most frequently studied, primarily derived from bone marrow, adipose tissue, or umbilical cord. MSCs were favoured due to their anti-inflammatory, anti-apoptotic, and pro-angiogenic properties, which may underlie the improvements in remodelling and functional status observed in studies by Bolli et al. (2021) [38], Qayyum et al. (2023) [31], and Bartunek et al. (2017) [39]. These types of cells have demonstrated significantly more efficacy mostly attributed to their paracrine and immunomodulatory effects.

In contrast, studies using bone marrow mononuclear cells (BM-MNCs) showed more variable results. These cells form a heterogeneous mix that includes hematopoietic and endothelial progenitor cells; however, they lack the purity and paracrine potency of mesenchymal stem cells (MSCs). Similarly, while trials using cardio-sphere-derived cells (CDCs), such as those by Bolli et al. (2021) [38], have shown promise, further research is needed to confirm their efficacy and feasibility in larger populations.

Beyond differences in cell type and source, substantial methodological heterogeneity exists in how cells are processed across studies. Variations in isolation techniques, culture expansion protocols, passage number, cryopreservation methods, and assessment of cell viability at the time of administration were inconsistently reported. These factors are known to influence cell potency, paracrine signalling, and immunomodulatory capacity, and may partly account for the variability in clinical outcomes observed between trials, even among studies using ostensibly similar cell populations.

Routes of administration, cell dose, and treatment timing

Several parameters alongside with the type of cells used influence the efficacy and safety of stem cell therapy in IHD and HF, particularly the route of administration, cell dosing, and timing. These factors are widely recognized in the literature as key modulators of clinical outcomes. Importantly, these parameters often varied simultaneously within individual trials, contributing to considerable methodological heterogeneity that limits direct comparison between studies and complicates interpretation of pooled treatment effects.

The intracoronary infusion was the most common route in our review, favored for its minimally invasive nature and procedural familiarity. However, it may be less effective in infarcted or hypoperfused regions due to limited cell homing.

More targeted methods, such as trans endocardial or intramyocardial injection, although invasive, may offer improved outcomes by delivering cells directly into viable myocardium. Trials by Perin et al. and Qayyum et al., using intramyocardial delivery, demonstrated notable improvements in LVEF, exercise capacity, and biomarkers, highlighting the potential of localized administration.

Intravenous delivery, while easier, was less commonly used and often less effective due to pulmonary cell trapping and poor cardiac homing. However, new strategies are being developed to address these limitations. For example, Hsiao et al. [40] reported early success using a combined intravenous and intracoronary approach, administered two days apart, which appeared safe and potentially more effective.

Cell doses varied widely across studies, typically ranging from 10⁶ to 10⁸ cells. Higher doses (> 10⁷) were more consistently linked with improvements in LVEF, 6MWT, and NT-proBNP. For instance, Attar et al. [29] delivered 20 million Wharton’s jelly-derived MSCs intracoronary, with significant functional benefits. In contrast, lower doses often produced marginal or no effects.

Evidence of a dose-response relationship is emerging, although the optimal range and safety ceiling remain unclear. Repeat dosing may further enhance efficacy. Gong et al. demonstrated that repeated IV infusions in rats with dilated cardiomyopathy resulted in better left ventricular (LV) function than a single dose, suggesting the importance of both dose size and schedule.

Timing also influences the therapeutic effect. Some trials administered cells within weeks of MI, while others targeted chronic HF. The post-MI myocardium, characterized by inflammation and remodelling, may be exceptionally responsive, but also prone to risks such as arrhythmia and poor cell survival. In contrast, chronic HF patients may benefit more from paracrine effects like neovascularization and reduced fibrosis. Gong et al. [41] further emphasized this by showing that subacute administration of umbilical cord-derived MSCs post-MI yielded better outcomes than delayed treatment, supporting the idea of a biological window of enhanced responsiveness that warrants further study.

Comparison with recent high-quality systematic reviews

Our findings are broadly consistent with previously published systematic reviews in this field, particularly the review by Abouzid et al. [42], which synthesized 35 studies published between 2007 and 2022, including RCTs, animal studies, and prior systematic reviews. In contrast, the present review was restricted to human randomized controlled trials and extended the search period from inception through August 2024, resulting in a more clinically relevant and methodologically homogeneous evidence base, despite residual inter-study heterogeneity.

While the overall conclusions are broadly aligned, with both reviews and previously published literature in reporting statistically significant but modest improvements in cardiac function, e.g., improvement in LVEF and reduction in LVEDV and LVESV following stem cell therapy, important methodological differences distinguish the current analysis.

The present study contributes additional value through several methodological refinements. These include and not limited to the uniform application of the Cochrane Risk of Bias 2.0 tool across all included trials with transparent reporting of individual bias domains, time-stratified meta-analyses conducted at predefined follow-up intervals with prespecified subgroup analyses by cell type and delivery route, and a comprehensive pooled quantitative synthesis of safety outcomes. Specifically, all-cause mortality, heart failure worsening, arrhythmic events, and hospitalization outcomes were systematically examined, providing a more structured assessment of safety than in prior reviews.

Overall, while reinforcing existing evidence that stem cell therapy is associated with modest improvements in cardiac structure and function in post-myocardial infarction and heart failure populations, the present findings offer a more granular and clinically contextualized interpretation of both efficacy and safety.

Broader context of cellular therapies and cardiovascular risk

From a broader translational perspective, evidence from cardio-oncology suggests that cellular and biologic therapies used in hematologic malignancies may confer late cardiovascular complications, even when early outcomes appear favorable. Long-term observational studies of patients with hematologic malignancies and survivors of hematopoietic stem cell transplantation have reported increased risks of heart failure, arrhythmias, cardiomyopathy, and early coronary artery disease years after treatment. This indicates persistent biological effects rather than transient toxicity [43, 44]. Importantly, these outcomes are influenced not only by therapy but also by traditional cardiovascular risk factors, including hypertension, smoking, diabetes, dyslipidemia, and advancing age, which may amplify susceptibility to cardiovascular injury. Further research is needed to clarify treatment-specific contributions to long-term cardiovascular risk and identify strategies for mitigation.

Research indicates that long-term changes in vascular health following hematopoietic stem cell transplantation are, in part, linked to ongoing damage to the endothelium caused by conditioning treatments and graft-versus-host disease. This contributes to accelerated vascular aging and a higher risk of cardiovascular issues later in life [45]. Additional findings from studies on vascular toxicity related to cancer therapy suggest that endothelial dysfunction and reduced regenerative capabilities may lead to prolonged vascular damage and a state that promotes atherosclerosis. This provides a biological explanation for the long-term cardiovascular complications seen in patients who have survived cancer [46].

Although stem cell therapy for ischemic heart disease and heart failure differ in therapeutic intent, these observations support a shared principle that the cardiovascular effects of cellular therapies are patient-dependent. In this context, the modest but consistent functional improvements observed in our analysis, together with substantial heterogeneity in clinical response, highlight the need for improved patient selection, biomarker-guided risk stratification, individualized risk assessment, and long-term safety surveillance, consistent with modern practices in cardio-oncology practice [47].

Limitations

This meta-analysis has several important limitations that warrant careful consideration. First, there was substantial clinical and methodological heterogeneity across the included randomized controlled trials. Patient populations differed widely by disease stage (acute or subacute myocardial infarction versus chronic heart failure), baseline left ventricular function, comorbidity burden, and use of guideline-directed medical therapy. In addition, there was considerable variability in stem cell characteristics. These included cell type, dose, processing protocols, route of administration, and timing of delivery. All of these factors could affect therapeutic efficacy and contribute to heterogeneity in the pooled analyses.

Second, although the Cochrane Risk of Bias 2.0 tool found generally low risk in areas such as outcome measurement and attrition, many studies showed high or unclear risk for selective outcome reporting. Such reporting can stress favorable echocardiographic endpoints and omit neutral or negative results, possibly inflating reported treatment effects. Funnel plot asymmetry for LVEF at certain follow-up periods also points to potential publication bias. Together, these biases may have changed the magnitude of pooled effect estimates and reduced the validity and reliability of some outcomes. Furthermore, although subgroup analyses were conducted to explore sources of heterogeneity based on stem cell type, route of administration, and follow-up duration, the limited number of trials within individual subgroups constrained the interpretability of these findings.

A formal meta-regression was not undertaken, in accordance with methodological recommendations, due to the small number of studies per covariate, substantial overlap between key study characteristics (including cell type, dose, and delivery route), and inconsistent reporting of critical methodological variables such as cell viability, processing protocols, and timing of administration. Under these conditions, meta-regression could yield unstable or misleading estimates, limiting its interpretive value.

Finally, several clinically meaningful endpoints, particularly patient-reported outcomes such as the Kansas City Cardiomyopathy Questionnaire (KCCQ) and the Minnesota Living with Heart Failure Questionnaire (MLHFQ), were inconsistently reported or underpowered. This reduced the ability to draw firm conclusions regarding patient-centered benefits. In addition, most trials employed short- to medium-term follow-up durations, which restricted the assessment of long-term efficacy and late safety outcomes. Such outcomes include arrhythmogenesis and myocardial fibrosis. Consequently, the generalizability of these findings to broader and more diverse cardiac populations remains limited.

Future directions and research priorities

This meta-analysis implies that in order to increase the efficacy of heart regeneration treatment, consistency and better standards of technique should be given top attention in future studies, together with the integration of emerging paradigms in precision cardiovascular medicine. Large-scale, multi-center randomized controlled trials and developed treatments controlling stem cell production, distribution, and administration should take priority, as future progress in stem cell therapy is likely to depend not only on further standardization of treatment protocols but also on advances in patient selection and risk stratification. Comparative research depends on well-stated clinical goals, among which functional ability, reverse remodeling, and constant improvement in left ventricular ejection fraction should be included. Mechanistic research enables the identification of paracrine signaling routes and cellular connections responsible for the observed functional advantages, while multi-omics approaches may facilitate the identification of biomarkers predictive of therapeutic response and adverse events. Especially when several stem cells are provided using creative or combination approaches, close attention should be paid to immunological properties and biodistribution kinetics. Further increasing efficacy involves investigating multimodal strategies, including repeated dosing, gene-modified cells, or bioengineered scaffolds. Advances in artificial intelligence may support more refined patient stratification and personalized treatment regimens, moving beyond a one-size-fits-all approach. Identifying patient subgroups most likely to benefit remains central to research, with treatment potentially tailored through customized timing following myocardial infarction or chronic heart failure, biomarker-driven stratification, and imaging-guided viability assessment. Treated groups must remain under continuous observation to assess potential carcinogenicity, late adverse effects, and long-term benefits. Lastly, standardized patient-reported outcomes integrated with health economics assessments are essential to determine the financial feasibility and real-world impact of stem cell therapy, while future studies should also evaluate the feasibility of applying this technology in resource-limited settings where ischemic heart disease is more prevalent and access to contemporary therapies is restricted.

Conclusion

This updated systematic review and meta-analysis support stem cell therapy as a safe and relatively successful additional treatment for HF and IHD patients. The intervention showed statistically significant increases in LVEF, decreases in LVEDV and LVESV, and improvements in specific clinical and quality-of-life measures, including NYHA class and MLHFQ scores in 35 randomized controlled trials. Among several types of cells, MSCs especially those derived from bone marrow, adipose tissue, and umbilical cord are the most beneficial. Routes of intramyocardial and intracoronary injection are linked to greater increases in heart function than those of intravenous methods. Generally speaking, the intervention was well tolerated and showed lower appreciable side effects compared to the placebo. Still, generalizability is limited in research by variations in cell preparation, dosage, and injection technique; inadequate long-term data also limits generalizability. The low degree of functional enhancement emphasizes the need for changing therapy parameters and improved patient choice processes. Though it is not now a replacement for conventional treatments, SCT shows promise as a regenerative approach in cardiovascular medicine. Its ultimate relevance in clinical practice has to be proven by more thorough, long-term research including longer follow-up.

Abbreviations

6MWT

6-Minute Walk Test

ABMMNCs

Autologous bone marrow mononuclear cells

ACE

Angiotensin-converting enzyme

AEs

Adverse events

AMI

Acute myocardial infarction

BM-MNCs

Bone marrow mononuclear cells

BMMNCs

Bone marrow mononuclear cells

β-blockers

Beta-blockers

CDCs

Cardiosphere-derived cells

CHF

Congestive heart failure

CI

Confidence interval

CMR

Cardiovascular magnetic resonance

CVDs

Cardiovascular diseases

EF%

Ejection fraction percentage

eGFR

Estimated glomerular filtration rate

GLS

Global longitudinal strain

HF

Heart failure

HFrEF

Heart failure with reduced ejection fraction

IHD

Ischemic heart disease

IV

Intravenous

KCCQ

Kansas City Cardiomyopathy Questionnaire

LDL-C

Low-density lipoprotein cholesterol

LV

Left ventricle

LVEDV

Left ventricular end-diastolic volume

LVEDVI

Left Ventricular End-Diastolic Volume Index

LVESV

Left ventricular end-systolic volume

LVESVI

Left Ventricular End-Systolic Volume Index

LVEF

Left ventricular ejection fraction

MACE

Major adverse cardiac events

MI

Myocardial infarction

MLHFQ

Minnesota Living with Heart Failure Questionnaire

MRA

Mineralocorticoid receptor antagonists

MSCs

Mesenchymal stem/stromal cells

NT-proBNP

N-terminal pro-B-type natriuretic peptide

NYHA

New York Heart Association

PAD

Peripheral arterial disease

PET

positron emission tomography

RCTs

Randomized controlled trials

SAEs

Serious adverse events

SGLT2

Sodium-glucose cotransporter-2

SCT

Stem cell therapy

SPECT

Single-photon emission computed tomography

Authors’ contributions

G.A.A had full access to all the data in the study and takes responsibility for its integrity and the accuracy of the data analysis. All authors contributed to data abstraction and drafting the first version of the manuscript. M.F.M. and F.A.A. supervised the study. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Data availability

All data generated or analyzed during this study is included in this published article.

Declarations

Ethics approval and consent to participate

This systematic review and meta-analysis did not involve original data from the participant sample; thus, ethical approval was not required. The study was registered in PROSPERO (CRD42024582716).

Consent for publication

Not applicable

Competing interests

The authors declare no conflicts of interest.