Introduction

Vascular diseases (VDs) are harmful alterations in the body’s vascular system, arising from damage or abnormalities in the blood vessels and failures within the blood vessel network. The major VDs include cerebral ischemia (CI) disease (e.g., stroke, hypoxic-ischemic encephalopathy [HIE]), cardiac ischemia disease (e.g., myocardial ischemia [MI]), and peripheral ischemia disease (e.g., peripheral artery disease). VDs affect people’s work and lives and are potentially life-threatening, bringing an inevitable burden on families and society1,2. To date, pharmacotherapy (e.g., thrombolytics, neuroprotective agents) and physiatrics (e.g., hypothermia therapy) have been widely used to treat VDs and, when necessary, combined with surgery3,4,5,6,7. However, thrombolytic drugs are only effective within 4.5 h of stroke onset8, and the mortality rate of HIE treated with hypothermia therapy is 25%9. A variety of surgical techniques have been developed for the treatment of atherosclerotic diseases, such as atherectomy, plain balloon angioplasty, drug-coated balloons, bare metal, and drug-eluting self-expanding stents10. However, these methods are not one-and-done, and over time, the blood vessels will narrow again11. Thus, pharmacotherapy and surgery only work within limited time windows and, even then, may lack effectiveness.

Cell-based therapy has recently gained growing attention as it represents a potentially safer and more effective therapeutic approach with an extended treatment window. Given their capacity for self-renewal and multilineage differentiation12, stem cells offer great potential for the regeneration and repair of damaged tissues. Although embryonic stem cells possess remarkable pluripotency, their clinical application remains limited due to ethical concerns13. Mesenchymal stem cells (MSCs), which can be isolated from various adult tissues, including adipose tissue or bone marrow14,15, exhibit multipotent differentiation15,16 and significant paracrine effects17,18. However, their procurement often requires invasive procedures, which pose additional risks and limit clinical applicability. In contrast, orally derived mesenchymal stem cells (OMSCs), including dental pulp stem cells (DPSCs) and stem cells from human exfoliated deciduous teeth (SHED), can be obtained non-invasively from discarded biological materials such as extracted wisdom teeth or exfoliated deciduous teeth. OMSCs possess comparable or even superior capacities for proliferation, differentiation, and secretion of bioactive factors19,20, making them an attractive and ethically acceptable source for regenerative therapy.

In recent years, OMSCs, their conditioned media (CM), and OMSC-derived exosomes have been investigated as novel treatment strategies for VDs in several animal studies21,22,23. These studies employed diverse administration routes, including intracranial, intranasal, and systemic delivery, and consistently demonstrated improvements in motor and sensory function. Notably, DPSCs have been shown to promote functional recovery in models of CI22, while SHED-derived CM provided superior neuroprotective effects compared with those of BMMSCs21. OMSCs and their secretome have been shown to exert these therapeutic effects mainly by promoting cell migration22, enhancing anti-inflammatory effects24, and inhibiting apoptosis25.

Although some systematic reviews have examined the effectiveness of stem cells in treating various VDs, research specifically on OMSCs as a unique subtype of MSCs has not been systematically or thoroughly analyzed. Therefore, this systematic review aimed to comprehensively evaluate the therapeutic potential of OMSCs and their derivatives in the treatment of VDs. The review sought to synthesize and analyze evidence from in vivo studies to determine the efficacy of various OMSC-based interventions, including cell transplantation, conditioned media, and exosome administration. In doing so, it compared outcomes across different OMSC sources and delivery routes, explored the mechanisms underlying vascular repair, and identified existing limitations in the current body of evidence. Ultimately, the objective was to establish an evidence base that could guide future research and support the clinical translation of OMSC-based therapies for VDs.

Methods

Protocol registration

The protocol was registered on the International Prospective Register of Systematic Reviews (PROSPERO) database (registration number CRD42022258794), which is available from: https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=258794. This review followed the PRISMA 2020 guidelines.

Search strategy

The literature search scope was from the inception of databases (PubMed, Embase, and Web of Science) to December 31, 2024. Search terms included combinations of “dental mesenchymal stem cells” and “vascular diseases” or their synonyms, as mentioned in Supplementary File 1. Two reviewers screened the literature using the same criteria, and when there were differing opinions, a third reviewer reviewed the literature and helped resolve any discrepancies.

Study selection

We included all studies reporting in vivo experiments using OMSCs (including side population cells derived from OMSCs), OMSC-CM, or OMSC-exosomes, regardless of the outcome, in animal models of VDs. Reviews, editorials, conference abstracts, case reports, meta-analyses, studies describing only in vitro data, non-English literature, and duplicate studies were excluded. Only animal studies were included in this systematic review.

Data extraction and quality assessment

Two reviewers independently extracted data using the same standards. Any disagreement was discussed with a third reviewer to resolve. Studies were classified based on the disease and injury models studied. The type of animal, source of OMSCs, VD model, administration route, treatment dose, and outcomes were recorded. Key parameters of the experimental design and methodology were extracted26 to identify whether the (1) species and strains of experimental animals were reported, (2) group and number of experimental animals were specified, (3) experimental animals were randomly assigned, and (4) investigators experimented in a single-anonymized manner. “The ARRIVE Essential 10: Compliance Questionnaire”27 was used to evaluate the included studies for bias and quality, and scored 18 subitems (Yes: 1; No: 0). The experimental outcomes were summarized qualitatively. The quantifiable indicators were extracted, taking into account the homogeneity of the study and the calculability of the effect size. Only CI studies that reported infarct volume in mm3 met this criterion; quantitative data were extracted from these studies, and a meta-analysis was conducted.

Statistical analysis

For each included study, the means and standard deviations of infarct volumes resulting from CI were extracted for quantitative synthesis. Meta-analysis was conducted using the inverse variance (IV) weighting method under a random-effects model, with between-study variance estimated using the DerSimonian and Laird method. Given the potential differences in measurement scales across studies, standardized mean differences (SMDs) were calculated as the summary effect size, facilitating comparability across diverse outcome metrics. Statistical significance was determined using a two-sided p-value threshold of < 0.05. Heterogeneity among studies was assessed using the I2 statistic, with values greater than 50% indicating substantial heterogeneity28. To evaluate the robustness of the results, sensitivity analyses were performed using a fixed-effects model. All meta-analyses were performed using Cochrane Review Manager software (RevMan Web, https://revman.cochrane.org/).

Results

Characteristics of selected studies

A total of 295 studies were obtained after a systematic search in three databases. After the elimination of duplicates, 221 records remained. In the title and abstract screening, 166 articles were excluded because they did not meet the inclusion criteria (unrelated to OMSCs or VDs, reviews/editorials/conference abstracts/notes/meta-analysis, and retraction). After the full-text screening of the remaining 55 records, 14 were excluded because they were either unrelated to or lacked data regarding animal work or the use of OMSCs to treat VDs. Thus, a total of 41 studies published between 2007 and 2024 were included in the systematic review (Fig. 1); In terms of cell type, DPSCs were the most frequently investigated (29 studies used DPSCs from human sources, one of which also studied human periodontal ligament stem cells [hPDLSCs], and three studies used DPSCs from rats), followed by SHED (8 studies), human gingival mesenchymal stem cells (hGMSCs) (1 study) and puppy deciduous teeth stem cells (1 study).

Fig. 1
Fig. 1
Full size image

PRISMA Flowchart.

We grouped the 41 studies into 5 categories based on different VDs: (1) CI (20 studies)21,22,23,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45, (2) HIE (5 studies)24,46,47,48,49, (3) MI (4 studies)25,50,51,52, (4) Hindlimb ischemia (HI) (7 studies)38,53,54,55,56,57,58, and (5) others (6 studies, including subarachnoid hemorrhage [2 studies]59,60, vascular dementia [1 study]61, atherosclerosis [2 studies]62,63, and degenerative valve disease [1 study]64. One study38, simultaneously reported on two animal models; therefore, it is included under both CI and HI.

Study quality and bias assessment (Table 1)

Table 1 Summary of the results and quality assessment of the studies.
Full size table

The bias and quality of the studies were evaluated based on “The ARRIVE Essential 10: Compliance Questionnaire”27. Supplementary file 2 summarizes the evaluation scores of the studies (Table S1) and the criteria used for evaluation (Table S2). In summary, 10 studies did not report the characterization of OMSCs. All studies reported the species of experimental animals, and only 1 study did not report the strain of the experimental animals54. One study38, simultaneously reported on two animal models (CI and HI). Random assignment of experimental animals was reported in 26 out of 41 studies (17 of 20 in CI, 2 of 5 in HIE, 2 of 4 in MI, 1 of 7 in HI, and 5 of 6 in others). More than half of the studies (21 of 41) reported the blinding of investigators or lab personnel (13 of 20 in CI, 4 of 5 in HIE, 1 of 4 in MI, 2 of 7 in HI, and 2 of 6 in others). Fifteen studies reported both randomization and blindness, accounting for more than one-third of the total. A summary of the study quality and bias assessment is shown in Table 1. For the studies that used OMSC-CM, the details of CM are summarized in Table S3 of Supplementary File 3.

CI studies (Table 2)

Table 2 Summary of CI studies.
Full size table

The studies of CI models can be divided into transient middle cerebral artery occlusion (T-MCAO) (15 studies), permanent middle cerebral artery occlusion (P-MCAO) (4 studies) and chronic CI (1 study). The T-MCAO models were established using the intraluminal thread to block the right or left middle cerebral artery and reperfusion after 2 h (8 of 15 studies), 1.5 h (5 of 15 studies), 1 h (1 of 15 studies), or 0.5 h (1 of 15 studies) of occlusion. In addition to the intraluminal thread, coagulation and cutting of the MCA and photothrombosis were also used in the P-MCAO model. A study on chronic CI used ligation of bilateral carotid arteries. All but two studies of T-MCAO and one study of P-MCAO used mice as the research model, while the others used rats. DPSCs were investigated in 17 studies; 13 used human DPSCs, while 4 used rat and porcine DPSCs. The other three studies utilized SHED and SHED-CM as the treatment group. Most studies used intracranial administration. Tail-vein injection, femoral vein cannula, intra-arterial, and intranasal administration were also reported. The cell injection amount ranged from 2*104 to 6*106 cells. The CM volume and exosome dosage were 100 µl SHED-CM, 10 µg total protein, and 100 µg exosome, respectively. The studies utilized various CI models, with transient middle cerebral artery occlusion (T-MCAO) being the most common, predominantly in mice. Treatment involved OMSCs, their CM, or exosomes, administered via multiple routes.

Nineteen of the 20 studies reported improvements in functional tests of motion or sensation. Fourteen studies demonstrated that transplantation of OMSCs (including exosomes and CM with OMSCs) significantly reduced infarct volume compared with the control group. Different studies also reported that OMSC-based transplantation reduced corpus callosum atrophy22, ameliorated cerebral edema23,36,37,45, and improved body weight43,44 in subject animals. Whether injected through the tail vein or intracranially, the transplanted OMSCs migrated to the ischemic area21,22,30,31,39,40,45, survived21,39,40, and were not found in systemic organs39. In another study42, in which hDPSCs were injected through the femoral vein, the cells were found to remain in the brain for 4 weeks. Seven studies described the effects of OMSC-based transplantation on neural regeneration, including differentiation into neurons and astrocytes22,39,41, reduction of neuron damage29,32,33, and expression of more neural markers21,30, and neurotrophic factors21,40. Four reports revealed that hDPSCs reduced microglial activation29,32 and astrogliosis39,42. In addition, OMSC treatment inhibited inflammation23,29,32 and apoptosis33,36,37,45. Only two studies32,39, in which hDPSCs were injected into the tail vein, reported the promotion of angiogenesis. Notably, Yew et al.42 reported that neither angiogenesis nor synaptogenesis in peri-infarct tissue was the primary cause of faster recovery in the hDPSC-treated group. Sugiyama, et al.35 found no evidence of differentiation of transplanted cells into neurons and endothelial cells; however, OMSCs enhanced the migration of neural progenitor cells (NPCs) from the subventricular zone (SVZ) to the peri-infarct region21,40. miR-877-3p was considered to be transferable through DPSC-extracellular vesicles (EVs), demonstrating neuroprotective effects and the potential to promote the survival and recovery of neurons after ischemia45. Finally, it was reported that hDPSC-exosomes inhibited I/R-mediated toll-like receptor 4 (TLR4) and nuclear factor kappa-B (NF-κB) activation, as well as high mobility group box 1 (HMGB1) cytoplasmic translocation23, and that hDPSCs reduced blood-brain barrier leakage and increased tight junction protein levels32.

Meta-analysis of infarct volume in CI

Based on the quantifiable indicators we extracted, only the CI studies reported infarct volume in mm³ provided adequate homogeneity and calculability of the effect size; therefore, only these studies were included in a meta-analysis. Two papers reported infarction volume as a percentage (unit: %), while another article reported infarction area as a percentage (unit: %). These were not included in the meta-analysis. Missing data were excluded, including reports that contained only the infarction ratio or lacked sample size information. Six studies were pooled21,29,32,35,39,40. Data on infarct volume in CI studies were pooled, and the studies were divided into two subgroups based on the study endpoint: short-term (within 10 days) and long-term (10 + days) follow-up. The p-value of the test for subgroup difference was 0.0006 (p < 0.05) (Fig. 2). If the subgroup effect was ignored, the overall heterogeneity was 40% (< 50%). As heterogeneity was reduced to 0% within each subgroup, this indicated that the duration of follow-up significantly influenced the treatment effect; both subgroups showed a significantly greater treatment effect than the control group (p-values < 0.00001). When using the fixed-effects model to investigate sensitivity (Fig. 3), the pooled standardized mean differences for the short- and long-term follow-up periods were − 1.56 (95% confidence interval: − 2.16, − 0.96) and − 5.23 (95% confidence interval: − 7.23, − 3.23), respectively, which were identical to that yielded by the random-effects model (since I2 = 0%). The overall (combination of the two groups) pooled standardized mean difference estimated by the fixed-effects model was − 1.87 (95% confidence interval: − 2.44, − 1.29), which was slightly lower than the difference of − 2.19 (95% confidence interval: − 3.01, − 1.37) estimated by the random-effects model. The funnel plot (Fig. 4) was slightly asymmetric, thus suggesting a low level of publication bias.

Fig. 2
Fig. 2
Full size image

Meta-analysis for the effect of OMSCs on infarct volume in CI models (random-effects model). If multiple sets of data were in the same study, these are represented by ordinal numbers (e.g., Nito 1 2018, Nito 2 2018). CI: cerebral ischemia; OMSCs: orally derived mesenchymal stem cells.

Fig. 3
Fig. 3
Full size image

Meta-analysis for the effect of OMSCs on infarct volume in CI models (fixed-effects model). If multiple sets of data were in the same study, these are represented by ordinal numbers (e.g., Nito 1 2018, Nito 2 2018). CI: cerebral ischemia; OMSCs: orally derived mesenchymal stem cells.

Fig. 4
Fig. 4
Full size image

Funnel plot for publication bias. SE: standard error; SMD: standard mean difference.

Across these studies, OMSC-based interventions consistently improved neurological function, reduced infarct volume and other structural brain injuries, and demonstrated cell migration to ischemic regions, neural protection and regeneration, modulation of glial responses, inflammation, and apoptosis, and, in a few studies, angiogenesis. Mechanistically, suggested pathways included enhanced NPC migration, transfer of neuroprotective miRNAs via extracellular vesicles, inhibition of TLR4/NF-κB signaling, and preservation of blood–brain barrier integrity. Meta-analysis of infarct volume data from six CI studies demonstrated a robust and significant reduction in infarct size with OMSC-based treatment in both short- and long-term follow-up, with low heterogeneity within subgroups and only slight evidence of publication bias, supporting a genuine and durable therapeutic effect.

HIE Studies (Table 3)

Table 3 Summary of HIE studies.
Full size table

The HIE disease model was used in 5 studies (4 used rats and 1 used mice). The surgical method was ligation or cauterization of the unilateral carotid artery and hypoxia treatment for 20 min, 1 h, and 2 h. All cells were from human sources (SHED, SHED-CM, and hDPSCs). The routes of administration included external jugular vein injection, subdural injection, intracardiac injection, and ventricular injection. The numbers of transplanted cells were 1*105, 2*105, and 1*106 cells, and the volume of the CM was 2 µl. All articles reported functional outcomes that improved following treatment. Three studies46,47,49 analyzed brain tissue damage after OMSC-based therapy, and only one showed that SHED and SHED-CM transplantation reduced brain tissue loss compared with the control group46. Three studies24,46,49 using hDPSCs showed that transplanted cells promoted the differentiation of neural lineages; one of these studies also reported that hDPSCs promoted angiogenesis and renal cortical blood flow24. Inhibition of apoptosis and inflammation was reported in all three studies. SHED and SHED-CM also improved survival rates and pathological scores46. Kitase et al. (2020) reported that SHED suppressed oxidative stress, decreased M1 microglia, and increased M2 microglia.

MI Studies (Table 4)

Table 4 Summary of MI studies.
Full size table

The four articles on MI used rats (permanent ligation of the left descending coronary artery/ anterior descending coronary artery), mice (ligation of the left anterior descending artery for 30 min), and rabbits (permanent ligation of the left anterior descending artery). When 1.5*106 hDPSCs were transplanted into rat myocardium, rat origin vessels increased, but did not differentiate into cardiac and smooth muscle cells50. Similarly, in a rabbit model, SHED injection increased neovascularization and myocardial cells, but their source was not explained51. Moreover, both studies reported a recovery of functional outcomes. SHED-CM (500 µl) inhibited apoptosis and inflammation in mice after injection through the jugular vein system25. hDPSC-EV treatment decreased the number of pro-inflammatory macrophages in the infarct area, thereby promoting the resolution of inflammation52. Reductions in infarct areas were reported in all studies.

HI Studies (Table 5)

Table 5 Summary of HI studies.
Full size table

Seven studies on HI were identified. Except for one study that reported that hDPSCs promoted the increase of CCL5 and VEGF expression54, the name of the ligated artery was not clearly specified. The remaining six studies all reported that animal models were established by femoral artery ligation. Approximately 70% of the studies (5 of 7) used blood flow imaging techniques and demonstrated that OMSC therapy enhanced blood reperfusion53,55,57,58. It was also reported that OMSCs improved functional outcomes (functional scoring and ischemia scoring)54,57,58 and reduced inflammation, fibrosis, and necrosis56,57. All seven studies utilized intramuscular injection and showed an increase in vascular structure.

Other VD Studies (Table 6)

Table 6 Summary of Other VD studies.
Full size table

Two studies utilized rDPSCs-CM59 and hDPSCs-exosomes60 for treating a rat model of subarachnoid hemorrhage, respectively. The decrease in Iba-1 positive cells demonstrated that rDPSCs-CM alleviated neuroinflammation59. Improvements in microcirculation and cognitive and motor impairments were also reported59. Exosomes obtained from hDPSCs could also alleviate neuroinflammation and inhibit microglial pyroptosis60.

One study described the treatment of rats with ligation of the bilateral common carotid artery using DPSCs from humans61. Following injection into the tail vein, cells (2 × 10^7 cells) migrated into the brain and improved neural marker expression (DCX, NeuN, and NF200) and functional outcomes (eight-arm maze test).

hGMSCs alleviated atherosclerosis in mice fed with high-fat food, which induced atherosclerosis via intravenous injection62. This study also observed that cell therapy reduced the expression of inflammatory monocytes (Ly-6Chi) and modulated cytokine expression (IFN-γ, IL-4, and IL-17 A). Similarly, hDPSCs decreased the ratio of neutrophils and monocytes and reduced inflammation63.

In the only included RCT64, dogs with chronic heart valve disease were used as research subjects, and strict enrollment criteria were established. Standard treatment of heart failure combined with intravenous puppy deciduous teeth stem cells (pOMSCs) was used as the experimental group, and standard treatment of heart failure alone was used as the control group. The results showed that the experimental group of dogs had improved heart function (including left ventricular ejection fraction) and American College of Veterinary Internal Medicine (ACVIM) functional class and quality of life scores.

Discussion

This systematic review of preclinical studies on the use of OMSCs for the treatment of VDs summarizes treatment evidence from several animal models of different VDs. The studies demonstrate that OMSC-based therapies are effective in animals with induced CI, HIE, MI, HI, subarachnoid hemorrhage, vascular dementia, and atherosclerosis in achieving favorable treatment outcomes. In these studies, OMSCs of different species (humans and rats) and tissues (DPSCs, SHED, PDLSCs, and GMSCs) were used in different forms (cells, CM, and exosomes) for treatment. These studies provide evidence for establishing disease models as standard modeling methods, as they are a prerequisite for evaluating whether OMSCs can effectively treat VDs. However, some articles did not report details on randomization and blinding, and in some cases lacked details on the number of samples.

While several systematic reviews have explored the potential of various stem cell types for treating VDs, our work is the first to focus exclusively on OMSCs. For instance, a previous systematic review by Nagpal et al.65 broadly examined the role of MSCs in stroke, including BMMSCs and adipose-derived stem cells (ADSCs), concluding that MSCs improve functional outcomes primarily through paracrine actions. However, that review did not differentiate between MSC sources, potentially overlooking the unique advantages of OMSCs, such as their neural crest origin and superior proliferative capacity. Similarly, Wang et al.66 conducted a meta-analysis on MSC therapy for myocardial infarction, reporting a significant reduction in infarct size. While informative, their analysis pooled data from BMMSCs, ADSCs, and umbilical cord-derived MSCs, introducing significant heterogeneity and failing to identify which cell type might be optimal. In contrast, our focused analysis of OMSCs, despite a smaller number of included studies for meta-analysis, provides a more homogeneous and specific evaluation of this cell source, revealing a clear signal of efficacy in CI models and highlighting its potential for other VDs for which quantitative synthesis was not yet possible.

Improvement in functional outcomes in experimental animals is a direct manifestation of physiological rehabilitation, as noted in the articles included in our study (Tables 2, 4, 5 and 6). Functional experiments were performed in 75.6% of the studies (31 of 41). The positive changes in local and systemic functions demonstrated the effectiveness of OMSC-based treatment. Although improvements in functional outcomes have been demonstrated in prior studies of stem cell therapy for VDs, the survival and differentiation of transplanted cells have been rarely reported67,68,69,70. Functional recovery by DPSCs after CI is most likely accomplished through DPSC-mediated paracrine effects rather than neural differentiation22. A similar proposed mechanism has also been reported for SHED46,49, which was observed to secrete paracrine factors that support and protect the brain microenvironment after HIE49; however, the cell differentiation status after transplantation was not reported. It is worth noting that differentiation and the paracrine effect are not mutually exclusive. In vascular dementia, implanted DPSCs can differentiate into the neural lineage and accelerate axon growth through the paracrine effect61.

At the mechanistic level, the effects of OMSCs include migration, differentiation, and alleviation of apoptosis and inflammation. DPSCs injected via the tail vein can cross the blood-brain barrier and survive and differentiate at the boundary of the ischemic region, ultimately reducing cerebral edema and infarct volume30. In HIE, DPSCs can migrate to the injury site and closely associate with the surrounding astrocytes without reducing tissue loss47. The injected OMSCs can differentiate into neural and endothelial cells, and SHED-CM is also believed to enhance the differentiation and migration of endogenous NPCs21. OMSC-related transplantation also protects against insult by reducing inflammation and apoptosis23,25,29,30,32,33,45,48,50,59,62,63. Li et al. reported for the first time that DPSC-exosomes can attenuate cerebral ischemic inflammation and provide protection by inhibiting the HMGB1/TLR4/Myeloid differentiation primary response 88 (MyD88)/NF-κB pathway23.

Notably, some studies have attempted to modify OMSCs. DPSCs combined with brain-derived neurotrophic factor (BDNF) may promote neuropathologic rehabilitation43. HGF not only can enhance the neuroprotective effect of DPSCs43, but also assist DPSCs in ameliorating atherosclerosis63. Compared with hDPSCs, anti34a-hDPSCs were more effective in inhibiting apoptosis, reducing cerebral edema and infarct volume, and improving motor function37. Other studies have reported the promotion of neuron and neurite regeneration by IGF1R+DPSCs24, as well as an increase in VEGF secretion due to HIF1-α overexpression, thereby enhancing recovery from HI54. Moreover, we found that there were five reports comparing the therapeutic effects of OMSCs with other MSCs. With the exception of one study from Moradi et al.54, the remaining four studies showed that OMSCs had a superior therapeutic effect to that of BMMSCs or ADSCs21,25,38,39. Additionally, when comparing OMSCs internally, it appears that SHED and PDLSCs are more effective than DPSCs.31,56. These findings suggest that there is still significant potential for research on OMSC-based treatment strategies in VDs. This comparative advantage of OMSCs over other MSC types is a critical finding that distinguishes our review. A meta-analysis on stem cell treatment for peripheral artery disease71 found that MSC therapy was effective; they did not find a significant difference between cell sources, likely due to the heterogeneity of the included studies and the lack of direct head-to-head comparisons. Our review, by specifically extracting data from studies that performed such direct comparisons, provides stronger evidence for the potential superiority of OMSCs, particularly in neuroprotection and angiogenesis. This highlights the importance of not just pooling all “MSCs” together but investigating the nuanced differences between tissue sources.

OMSCs show promising therapeutic potential as a novel treatment for VDs. However, the key to treatment success in animal studies may depend on the type of animal model used, the administered dose, and the timing of intervention. For example, while stroke has different levels of severity and the best window for rescue, the majority of animal experiments have been limited to the investigation of a specific disease stage. In contrast, clinical treatment in humans is much more complex. In addition, unlike rapidly life-threatening diseases such as cancer, many patients with VDs can survive for a long time despite varying degrees of organ damage. Further advances in the clinical application of stem cell therapy may also be hindered by patient hesitancy to enroll in clinical trials and government restrictions due to ethical and safety issues. Unlike traditional small-molecule drugs, exogenous cells can survive, proliferate, and differentiate in the body for a longer period. For stem cell transplantation, there is a risk of tumor formation, which is a long-term, low-probability but fatal risk. This makes it difficult to completely eliminate potential hazards in preclinical animal experiments. Although current clinical studies on the use of stem cells to treat VDs have not yet involved OMSCs, early clinical trials of intravenous autologous BMMSCs transplantation in stroke patients have shown safe and effective results72. This makes the future clinical prospects of OMSCs worthy of anticipation. In addition, it is worth emphasizing that the majority of studies included in this systematic review employed whole-cell-based therapeutics, while a few studies utilized OMSC-based CM and EVs (including exosomes). These two treatment approaches, whole stem cell versus secretome therapies (e.g., CM and EVs), could differ in terms of risks, regulatory considerations, and pretreatment requirements. However, the emerging evidence also suggests that whole mesenchymal cell therapies also exert their effects mainly through their secretome rather than through differentiation73.

One limitation of this systematic review was that only 6 studies met the criteria for inclusion in the meta-analysis. Although having a larger number of studies can improve the statistical power and robustness of a meta-analysis, high-quality analyses can still be conducted with fewer studies if they are comprehensive and methodologically sound. The key considerations are the quality, relevance, and availability of studies suitable for pooling rather than just the number of studies. Emphasizing rigorous study selection and data synthesis is more important than reaching a specific numerical threshold. Scientific evidence suggests that, in practice, at least 5 studies are necessary to achieve greater statistical power than individual studies within a random-effects meta-analysis71.

Approximately half of the included articles did not have a clear blinding method. The core purpose of blinding is to ensure that the expectancy effects are evenly distributed between the treatment group and the control group. When blinding fails, this balance is disrupted, thereby creating “activated expectancy bias”. Apart from the introduction of blinding, there are also aspects such as the determination of sample size and the method of randomization, which may all reduce the quality of the studies, ultimately introducing heterogeneity, increasing bias, and even misleading the interpretation of the results. This risk of bias across the included studies may have led to an overestimation of OMSC efficacy. While 75.6% of studies reported functional outcomes, only 15 studies (36%; 15/41) reported combined randomization and blinding. The lack of clear blinding methods is particularly concerning for functional outcome scores, as it can introduce expectation bias where the investigators’ preconceptions may potentially result in higher scores of improvement. Consequently, the significant effect sizes observed in our meta-analysis must be interpreted with caution, as incomplete methodological rigor across the included studies may have inflated the therapeutic benefits.

Another limitation was potential publication bias; many unpublished preclinical studies may have used OMSCs to treat VDs but have yielded non-significant results. A limitation of the included studies was that they did not explore the possible adverse events of OMSC transplantation in depth; these remain to be determined by further studies before proceeding to clinical trials. Clinical treatment strategies involving OMSCs are still in their infancy, and evaluating their safety and effectiveness in humans is a priority.

Methodological heterogeneity remains a significant barrier to the clinical translation of MSC-based therapies. Across the included studies, we observed substantial variability in cell sources (e.g., hDPSCs, SHED, PDLSCs, GMSCs, etc.), dosages (ranging from 2*104 to 2*107 cells), and delivery routes (including intracranial, intranasal, intramuscular, arterial, and venous injections). Furthermore, the timing of intervention and the follow-up duration varied significantly, and the meta-analysis results confirmed that follow-up time significantly influenced the treatment effect size (p = 0.0006). Such diversity in experimental design limits cross-study comparability and underscores the need for standardized preclinical protocols to confirm effectiveness before advancement to human trials.

To the best of our knowledge, this is the first systematic review to focus on the therapeutic effects of OMSCs and their secretome in in vivo VD models. OMSCs hold considerable translational potential in the treatment of VDs due to their accessibility, ethical acceptability, and robust regenerative properties, positioning them as a promising alternative to MSCs obtained from bone marrow or adipose tissue. OMSCs exhibit potent angiogenic, anti-inflammatory, and anti-apoptotic effects, primarily mediated through their secretome, a complex mixture of cytokines, growth factors, extracellular vesicles, and exosomes. This paracrine signalling can promote tissue repair, enhance neovascularization, and modulate immune responses, all of which are essential for restoring vascular integrity following ischemic or inflammatory injury. The OMSC secretome, in particular, has emerged as a feasible cell-free therapeutic option that overcomes many limitations associated with direct stem cell transplantation, such as immune rejection, limited cell survival, and tumorigenic risks. Such findings suggest that OMSC-derived factors could be developed into standardized biologic formulations suitable for clinical use, offering reproducible and safe treatment modalities for patients with cerebrovascular, peripheral, or ischemic vascular conditions.

Looking ahead, future studies should focus on several critical areas to bridge the gap between preclinical evidence and clinical application. First, standardized protocols for OMSC isolation, culture, and characterization are needed to ensure reproducibility and comparability across studies. Second, the molecular composition and mechanisms of action of the OMSC secretome require in-depth profiling using advanced omics approaches to identify key therapeutic effectors. Third, optimized methods for exosome purification, storage, and delivery must be developed to ensure stability and bioactivity for clinical use. Additionally, rigorous dose–response studies, long-term safety evaluations, and large-animal models are essential to establish efficacy and minimize translational risks. Finally, well-designed early-phase clinical trials should be initiated to validate preclinical outcomes and define the therapeutic window, delivery route, and target patient populations most likely to benefit from OMSC-based interventions. By addressing these research priorities, OMSCs and their secretome have the potential to evolve from promising experimental therapies into clinically viable interventions for the treatment and prevention of vascular diseases.