The literature search yielded a total of 1202 potentially relevant publications. Among them, 462 studies were eliminated due to duplication. After the title and abstract review, 687 articles were excluded. After reading the full texts of 53 research articles, 27 articles were excluded due to ineligibility (see the supplementary materials for specific reasons), and 26 studies were finally included in this meta-analysis [25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50]. The literature screening flow chart is shown in Fig. 1. The citation track map in Web of Science database depicts the citation between important works of the research field. The earliest study contributing to the field emerged in 2006, which defined minimum standards for human mesenchymal stromal cells. (Fig. 2A). This field experienced an outburst between the year 2015–2020. We also found that in the PubMed database, published documents related to EVs and IUA are gradually increasing, showing that research in this field has developed rapidly and is in a rapidly rising stage. (Fig. 2B).

Bibliometric analysis. (A) Citation track map. Each circle represents documents (including DOI numbers) that have been cited more than 20 times, and are defined as representative documents in the field. The larger the circle, the more times it has been cited. (B) Analysis ofannual publishing trends in PubMed database. Abbreviation: DOI: digital object unique identifier

Table 1 summarizes the characteristics of the included studies. These 26 studies were conducted from 2018 to 2025 (Fig. 3A), of which 23 trials were conducted in China [26, 28,29,30,31,32,33,34,35,36,37,38, 40, 41, 44, 46,47,48,49, 51], 1 was conducted in Iran [39], 1 was conducted in Turkey [27], and 1 was conducted in Egypt [25] (Fig. 3B). The total sample size was 899 animals. 21 studies used rat animal models [25,26,27, 29,30,31,32, 34,35,36, 38,39,40,41,42,43,44,45, 48, 49], 5 studies used mice [28, 33, 46, 47, 50], and 1 study used both mice and rats [33] (Fig. 3C). Methods used for disease modeling include mechanic damage and chemical damage. Chemical damage includes trichloroacetic acid, ethanol, and lipopolysaccharides (Fig. 3D). The EVs used in the experiment were derived from various stem cells from humans, rats/mice or rabbits, including umbilical cord mesenchymal stromal cells (UCMSCs), bone marrow mesenchymal stem cells (BMSCs) and adipose tissue mesenchymal stromal cells (ADSCs). Although EVs were derived from cells of different species, each study independently demonstrated their effectiveness against IUA in animal models. The route of administration in the trial was mainly intravenous, but intrauterine and intraperitoneal injections were also used (Fig. 3E).

An overview of studies features. Distribution of (A) publications by year (B) region (C) animal model (D) disease model (E) route of administration

The risk of bias of the included articles is summarized in Fig. 4. Four studies specified that the animal groups were randomly generated [39, 47, 48, 50], and most studies reported baseline characteristics, but none described allocation concealment methods, so the risk of selection bias is unclear. Most of these studies discussed the issues of random housing and blinding of personnel. But obviously, the animals don’t know whether they are in the experimental or control group; therefore, the “participants” are blind. The overall risk of performance bias and reporting bias is low. Three study had incomplete outcome data and therefore had a high risk of attrition bias [35, 49, 50]. Overall, the risk of bias from the animal experiments included is unclear.

After endometrial basal layer injury, endometrial stem cells decrease, repair capacity decreases, and fibrosis increases. Endometrial thickness is an important indicator for assessing the severity of IUA and postoperative recovery and is highly correlated with pregnancy outcome. The endometrium of patients with IUA is usually thin, with disrupted or absent endometrium visible on ultrasound. An analysis including 17 studies reported the efficacy of stem cell-derived EVs on endometrial thickness [29, 30, 32, 33, 35,36,37,38,39,40, 43,44,45,46,47,48, 50]. Stem cell-derived EVs can significantly increase endometrial thickness (SMD = 2.65; 95% CI: 1.90 ~ 3.40; P < 0.00001; I2 = 56%, P = 0.0003) (Fig. 5A).

Forest plots depicting the comparison between the stem cell -derived extracellular vesicles and control groups: (A) Endometrial thickness; (B) Number of endometrial glands; (C) Fibrosis area; (D) Pregnancy rates; (E) Number of embryos

The number of glands is an important histologic indicator of endothelial function and influences the responsiveness to hormone therapy. Reduced glands lead to decreased endothelial secretion, which affects embryo implantation. Endothelial glands can be seen sparsely or absent on hysteroscopy in patients with IUA, and in severe cases, “scar-like” changes are seen. 17 studies reported the effects of stem cell-derived EVs on the number of endometrial glands [25, 28,29,30, 32, 33, 38,39,40, 42, 43, 45, 47, 48, 50, 51]. Meta-analysis results indicate that administration of stem cell -derived EVs can significantly increase the number of endometrial glands (SMD = 3.78; 95% CI: 2.62 ~ 4.93; P < 0.00001; I2 = 71%, P < 0.00001) (Fig. 5B).

Excessive deposition of collagen fibers (especially Col-1) in IUA replaces the normal endothelial mesenchyme and leads to fibrosis, thus affecting embryo implantation. 16 studies involving 140 animals were included to analyze the effects of stem cell-derived EVs on fibrosis area [25, 27, 29, 30,