Abstract
Background:
Accumulating evidence suggests that astragalus polysaccharide (APS) may enhance the efficacy of conventional cancer therapies through multiple mechanisms. However, the synergistic effects of APS have not been systematically quantified. This meta-analysis was therefore conducted to quantify these potential synergistic antitumor effects and provide preclinical evidence to inform future clinical trials.
Methods:
Following PRISMA 2020 guidelines, we systematically searched ten databases (including PubMed and Web of Science) for preclinical studies from inception to May 2025 using predefined inclusion criteria. Risk of bias was assessed using SYRCLE’s RoB tool. Meta-analyses and subgroup analyses were performed using RevMan 5.4.1, while publication bias was assessed via funnel plots and Egger’s test (Stata 17.0). This systematic review was prospectively registered in PROSPERO (registration number: CRD420251047751).
Results:
Forty-one publications (44 independent studies) involving 748 animals were included. APS combination therapy was associated with significant improvements in tumor-related outcomes, including reduced tumor weight and volume, suppressed metastasis, and prolonged survival. Mechanistically, APS co-administration enhanced CD8+ T-cell infiltration, increased splenic and thymic indices, modulated cytokine profiles (TNF-α, IL-2, IFN-γ, IL-12, IL-6, IL-10), and reduced PD-1/PD-L1 expression in tumor tissue. Additionally, APS appeared to alleviate chemotherapy-induced nephrotoxicity, as evidenced by lower serum creatinine levels. Subgroup analyses indicated that heterogeneity was partially explained by model type, APS dosing regimen, and combination therapy modality. The certainty of evidence for primary outcomes was rated as low or very low according to the GRADE assessment.
Conclusion:
This meta-analysis provides preclinical evidence that APS may serve as an adjunctive agent to enhance the efficacy of conventional cancer therapies. However, given the low certainty of current evidence, further mechanistic studies and well-designed clinical trials are urgently warranted to establish its efficacy and therapeutic role in oncology.
Systematic Review Registration:
PROSPERO 2025 CRD420251047751 https://www.crd.york.ac.uk/PROSPERO/view/CRD420251047751
1 Introduction
The clinical management of solid tumors increasingly relies on multimodal, precision-medicine-based strategies that integrate surgery, chemotherapy, radiotherapy, targeted therapy, and immunotherapy (National Cancer Institute, 2017; Wagle et al., 2025). Despite significant therapeutic advances, treatment outcomes remain limited by several persistent challenges, including an immunosuppressive tumor microenvironment (TME), intrinsic or acquired drug resistance, and dose-limiting toxicities that compromise treatment continuity and durable responses (Jairam et al., 2019; Liu B. et al., 2024). These limitations have driven increasing interest in adjunctive approaches capable of enhancing antitumor efficacy while mitigating systemic toxicity. Traditional Chinese Medicine (TCM) has attracted growing attention in this context due to its broad immunomodulatory activity and generally favorable safety profile. TCM applies the Fuzheng Quxie (扶正祛邪) principle to cancer therapeutics — a paradigm aimed at reinforcing host defense mechanisms (Zhengqi, 正气) while eliminating pathogenic factors (Xieqi, 邪气). This paradigm posits that effective tumor control requires concurrently enhancing physiological resilience to potentiate the effects of conventional anticancer therapies.
Astragalus mongholicus Bunge (AM; Fabaceae), known medicinally as Astragali Radix (Huangqi), is a foundational herb in TCM with documented use tracing back to the Han dynasty in the Shennong Bencao Jing (ca. 200 CE). For millennia, it has been employed to strengthen host defenses by tonifying qi and nourishing blood. Phytochemical studies have identified numerous bioactive constituents in Astragali Radix, primarily including polysaccharides, triterpene saponins (astragalosides), flavonoids, and trace alkaloids (Durazzo et al., 2021; Liu Y. X. et al., 2024). Among these, astragalus polysaccharide (APS) is regarded as the principal immunologically active component linking traditional tonifying functions with modern immune regulation research.
Since the 1970s, APS has been shown to modulate humoral and cellular immunity and promote hematopoietic recovery. In 2001, an injectable APS formulation (China NMPA approval Z20040086) was approved for managing chemotherapy-induced leukopenia and immunodeficiency.
Recent mechanistic studies further indicate that APS enhances antitumor immunity by stimulating immune activity in central and peripheral lymphoid organs; activating macrophages, natural killer cells, dendritic cells, and T lymphocytes; and inducing the secretion of antitumor cytokines (Li et al., 2022; An et al., 2022; Zhou L. et al., 2017). APS can also partially reprogram the immunosuppressive tumor microenvironment through multiple pathways, thereby restoring CD8+ T-cell effector function and strengthening immune surveillance (Li et al., 2020; Du et al., 2022). These immunomodulatory actions appear to directly target core pathophysiological barriers that limit the efficacy of cancer treatment. Existing clinical studies on APS injection have primarily focused on improvements in cancer-related fatigue (Chen et al., 2012; Wang et al., 2019; Shen et al., 2024; Chang et al., 2025), quality of life (Guo et al., 2012), and select immune parameters (Tsao et al., 2021), whereas few have examined its potential synergistic antitumor effects. Consequently, despite its biological plausibility and early clinical signals, APS has not been broadly integrated into routine oncology practice. This translational gap likely arises from fragmented efficacy data, inadequate cross-model validation, and insufficient mechanistic exploration of its synergistic potential with conventional therapies.
To help bridge this gap, we performed a systematic review and meta-analysis of preclinical studies to: 1) quantify the potential synergistic antitumor effects of APS when combined with conventional therapies; 2) delineate underlying molecular mechanisms through biomarker correlation analysis; and 3) evaluate methodological rigor using SYRCLE’s risk-of-bias tool to inform future preclinical and translational research.
2 Methods
This systematic review and meta-analysis was conducted in accordance with the PRISMA 2020 guidelines (Page et al., 2021) and was prospectively registered in PROSPERO (CRD420251047751).
2.1 Search strategy
A systematic literature search was conducted across ten electronic databases, including PubMed, Web of Science, Embase, Scopus, ProQuest, Ovid, the China National Knowledge Infrastructure (CNKI), Wanfang Data, VIP, and the China Biology Medicine Disc (CBM). The search timeframe spanned from database inception to May 2025, and the search was restricted to publications in Chinese or English. Medical Subject Headings (MeSH) combined with free-text keywords were used in the English-language databases. The search strategy incorporated the following key terms: 1) Problem: Neoplasms [MeSH], tumor, cancer, carcinoma, malignan*; 2) Population: animal model, animal experiment, mice, mouse, xenograft, tumor-bearing, transplantable tumor; and 3) Intervention: Astragalus polysaccharide, Astragalus polysacharin, APS. In addition, the Chinese databases were searched using the corresponding Chinese terms to ensure comprehensive retrieval of relevant studies.
2.2 Inclusion/exclusion criteria
2.2.1 Inclusion criteria
Population: Animal models bearing malignant neoplasms.
Intervention: Experimental groups administered APS in combination with conventional antitumor therapies.
Comparison: Control groups receiving conventional antitumor therapy alone.
Outcomes: Studies reporting at least one quantitative outcome related to antitumor efficacy (e.g., tumor weight, tumor volume, survival time, or mechanism-related biomarkers).
Study design: Randomized controlled animal studies.
Data availability: Sufficient quantitative data to calculate effect sizes (e.g., mean ± SD, sample size, p-values).
Language: Studies published in Chinese or English.
2.2.2 Exclusion criteria
Population: Clinical studies, in vitro experiments, or animal models of non-cancer diseases.
Intervention: APS monotherapy; crude Astragalus extracts lacking polysaccharide isolation; or multi-herbal formulations containing APS.
Control: Studies using no-treatment controls or non-standard antitumor therapies.
Outcomes: Studies without quantifiable antitumor outcomes or those reporting only qualitative findings.
Study design: Non-original research (e.g., abstracts, reviews, case reports); studies lacking an independent control group; or cross-over designs.
Data limitations: Insufficient quantitative data for meta-analysis (e.g., missing means, SDs, or sample sizes).
Language: Studies published in languages other than Chinese or English without accessible translations.
Other: Duplicate publications or studies with overlapping datasets.
2.3 Study selection and data extraction
The study selection process was conducted in two consecutive stages. Initially, two independent reviewers screened the titles and abstracts of all retrieved records for potential eligibility according to predefined inclusion criteria. In the second stage, the same reviewers independently evaluated the full texts of studies that passed the initial screening to determine final eligibility. Any disagreements were resolved through discussion with a third, senior reviewer. A standardized data-extraction form (Excel) was used to record key study characteristics, including intervention details, experimental design, and outcome measures. To ensure objectivity, a predefined decision rule was applied: when multiple measurements were available, the final time point and the highest APS dose were extracted. This a priori strategy was used to assess the maximum potential intervention effect, maintain consistency, and minimize post hoc selection bias. Numerical data were extracted directly from tables or text; when outcomes were presented only in graphical form, data were digitized using GetData Graph Digitizer software.
2.4 Quality assessment
The quality of the included animal studies was assessed using the SYRCLE Risk of Bias (RoB) tool. This tool evaluates ten methodological domains covering six types of bias (e.g., sequence generation, baseline characteristics, blinding). Each domain was rated as “L” (low risk), “H” (high risk), or “U” (unclear risk due to insufficient methodological reporting). Two reviewers independently conducted the assessment, and any disagreements were resolved by consensus with a third reviewer.
2.5 Statistical analysis
Meta-analyses were conducted using Review Manager (RevMan) 5.4.1. All analyzed outcomes were treated as continuous variables. Accordingly, we calculated mean differences (MDs) or standardized mean differences (SMDs) with 95% confidence intervals (CIs), selecting the effect size metric based on measurement scale consistency across studies. Heterogeneity was assessed using the I2 statistic and Cochran’s Q test, with I2 > 50% or Q-test p < 0.10 indicating substantial heterogeneity. When heterogeneity was not statistically significant (I2 ≤ 50% and p ≥ 0.10), a fixed-effects model was applied; otherwise, a random-effects model was used. Subgroup analyses and meta-regression were performed to explore potential sources of heterogeneity. Publication bias was assessed using funnel plot symmetry and Egger’s regression test (p < 0.05 considered significant) in Stata 17.0.
2.6 Sensitivity analysis
Sensitivity analyses were conducted using the leave-one-out approach, which involved sequentially omitting individual studies to assess the robustness of the pooled estimates. This approach quantified the influence of each study on the overall effect size.
2.7 Assessment of publication bias and evidence quality
Publication bias was assessed using funnel plot symmetry and Egger’s regression test in Stata 17.0, and only when ten or more studies were available for a given outcome. When publication bias was indicated (Egger’s test p < 0.05), the trim-and-fill method was applied to adjust for potentially missing studies. The difference between the original and trim-and-fill–adjusted effect sizes was calculated to quantify its impact on result robustness. The overall evidence quality was evaluated using the GRADE framework adapted for preclinical studies, considering risk of bias, inconsistency, indirectness, imprecision, and publication bias. Final certainty ratings were categorized as “High,” “Moderate,” “Low,” or “Very Low.”
3 Results
3.1 Study inclusion
A total of 2,259 records were identified through database searches. After systematic screening, 41 articles (reporting 44 independent experiments) met the predefined eligibility criteria and were included in the meta-analysis. Figure 1 presents the PRISMA flow diagram illustrating the study selection process.
FIGURE 1
3.2 Characteristics of the included preclinical studies
As detailed in Table 1, the 44 included preclinical studies investigated 11 distinct cancer typesLung cancer (n = 15, 34.1%) and liver cancer (n = 9, 20.5%) were predominant, followed by melanoma (n = 5, 11.4%), breast cancer (n = 4, 9.1%), cervical cancer (n = 3, 6.8%), and colorectal cancer (n = 3, 6.8%). Five additional types—bladder cancer, nasopharyngeal carcinoma, glioma, sarcoma, and pancreatic cancer—were each represented by single studies (n = 1 per type; collectively 11.4%).
TABLE 1
| Study | Cancer models | Cell strain and count | Animal species (sex/age or weight) | Sample size | Combination group | APS dosage/administration/frequency | Monotherapy group (dosage/administration/frequency) | Treatment initiation | Followup (days) | Outcome indicator |
|---|---|---|---|---|---|---|---|---|---|---|
| Huang et al. (2002) | Melanoma | B16, 5 × 106 | C57BL/6,F/M, 6–8 w,18–23g | 10,10 | APS + CD3AK | 50 mg/kg, i.p., qd ×3 | 5 × 106CD3AK cells/mouse, i.p., qd ×3 | 24 h post-modeling | 20 | ③⑥⑤ |
| Huo (2016) | Cervical | Hela, NR | BALB/c, F, 6–8 w,16–20g | 10,10 | APS + CIK | 200 mg/kg, v., biw ×6 | 5 × 107 CIK cells/mouse, i.v., biw ×6 | 6 days post-modeling | 21 | ① |
| Zhang et al. (2024) | Liver | Huh7, 5 × 106 | NOD/SCID, F/M, 6–8 w | 5,5 | APS + CAR-T | 50 mg/kg, p.o., qd ×21 | 5 × 106 CAR-T cells/mouse, i.v., qd ×1 | 5 days post-modeling | 28 | ①②⑤⑥ |
| Liver | HepG2, 2 × 106 | NOD/SCID, F/M, 6–8 w | 5,5 | APS + CAR-T | 50 mg/kg, p.o., qd ×21 | 5 × 106 CAR-T cells/mouse, i.v., qd ×1 | 5 days post-modeling | 28 | ①②⑥ | |
| Sha et al. (2022) | Melanoma | B16, 6 × 105 | C57BL/6, F, 6 w | 6,6 | APS + anti-PD-1 | 100 mg/kg, p.o., qd ×7 | 20 mg/kg, i.p., qod ×4 | When tumor volume reached 100 mm3 | 7 | ①②⑥ |
| Lin et al. (2024) | Liver | H22, 3 × 106 | BALB/c, M, 6–8 w,18–22g | 6,6 | APS + anti-PD-1 | 8 mg/mouse, p.o., qd ×14 | 200 ug/mouse, i.p., qd ×14 | When tumor volume reached 50–100 mm3 | 14 | ①②⑤⑥⑬ |
| Wang (2024) | Lung | LLC, 2 × 106 | C57BL/6, M, 6–8 w,18–22g | 5,5 | APS + anti-TIGIT | 100 mg/kg, i.p., qd ×14 | 5 mg/kg, i.p., qod ×7 | 7 days post-modeling | 14 | ①②⑥⑬ |
| Liu et al. (2024b) | Lung | LLC, 2 × 106 | C57BL/6, M, 6–8 w | 10,10 | APS + anti-PD-L1 | 80 mg/kg, i.p., qd ×14 | 200 ug/mouse, i.p., q4d ×4 | 2 days post-modeling | 14 | ①②④⑤⑥ |
| Lin et al. (2022) | Lung | LLC, 2 × 106 | C57BL/6, M, 18–22g | 10,10 | APS + anti-PD-L1 | 80 mg/kg, p.o.,qd ×14 | 200 ug/mouse, i.p., q4d ×4 | 7 days post-modeling | 14 | ②④⑤⑥ |
| Li et al. (2024) | Melanoma | B16, 1 × 105 | C57BL/6N,M, 6–8 w,18–22g | 5,5 | APS-NVs | 6.25 mg/kg, i.v., q4d ×3 | 25 mg/kg, i.v., q4d ×3 | 3 days post-modeling | 16 | ①②⑤⑥⑩⑪ |
| Lung | LLC, 1 × 105 | C57BL/6N,M, 6–8 w,18–22g | 6,6 | APS-NVs@L-Ag | i.v., q4d ×3 | i.v., q4d ×3 | 3 days post-modeling | 16 | ①②⑩⑪ | |
| Cao et al. (2024) | Colon | MC38, NR | C57BL/6, F, 6–8 w | 4,4 | NP-TCL@APS | 100 ul/mouse, s.c.,qw ×3 | 100 ul/mouse, s.c.,qw ×3 | 5 days post-modeling | 28 | ①②⑥ |
| Lv et al. (2017) | Liver | H22, 4 × 105 | KM mice,M, 6 w,15–25g | 10,10 | APS + RT | 200 mg/kg, i.p., qd ×14 | 5 Gy qod ×7 | 7 days post-modeling | 14 | ①③④ |
| Zhang (2019) | Lung | LLC, 2 × 106 | C57BL/6J,M, 6–8 w,18–22g | 15,15 | APS + RT | 174 mg/kg, p.o., qd ×7 | 8Gy qd ×1 | 9–12 days post-modeling | 13 | ①⑤⑥⑮ |
| Mu et al. (2019) | Liver | H22, 1 × 106 | KM mice, M, 8 w | 10,10 | APS + RT | 100 mg/kg, i.p., qd ×10 | 4Gy qd ×1 | When tumor volume reached 100 mm3 | 10 | ①②④⑤⑮ |
| Song et al. (2022) | Cervical | U14, 1 × 106 | KM mice, F, 18–22g | 10,10 | APS + RT | 2 mg/mouse, i.p., qd ×10 | 4Gy qd ×1 | When tumor volume reached 100 mm3 | 10 | ②⑫ |
| Huang et al. (2008) | Liver | H22, 1 × 106 | KM mice, F/M, 6–8 w,18–22g | 10,10 | APS + DDP | 200 mg/kg, i.p., qd ×18 | 2 mg/kg, i.p., q3d ×6 | 7 days post-modeling | 19 | ① |
| Tian (2012) | Liver | H22, 5 × 106 | KM mice, M, 6–12 w,18–22g | 10,10 | APS + DOX | 200 mg/kg, i.p., qd ×10 | 1.25 mg/kg, i.p., qod ×3 | When tumor volume reached 100 mm3 | 10 | ①④⑤⑭ |
| Sun et al. (2013) | Sarcoma | S180, 2 × 105 | BALB/c, M, 20–22g | 10,10 | APS + DDP | 300 mg/kg, p.o., qd ×14 | 2 mg/kg, i.p., qod ×7 | 2 days post-modeling | 14 | ①④⑩ |
| Ming et al. (2014) | Lung | LLC, 2 × 106 | C57BL/6, M, 6–8 w,18–22g | 10,10 | APS + DDP | 100 mg/kg, i.p., qd ×20 | 6 mg/kg, i.p., qw ×3 | 2 days post-modeling | 21 | ①④⑤⑫ |
| Zhuang et al. (2017) | Lung | LLC, 2 × 106 | C57BL/6J, M, 16–18g | 10,10 | APS + DDP | 60 ug/mouse, i.p., qd ×20 | 6 mg/kg, i.p., qw ×3 | 2 days post-modeling | 21 | ① |
| Wang et al. (2017b) | Lung | LLC, 2 × 106 | C57BL/6J, M, 8 w,18–22g | 10,10 | APS + DDP | 200 mg/kg, i.p., qw ×3 | 2 mg/kg, i.p., qw ×3 | When the tumor diameter reached 5 mm | 21 | ①⑥⑮ |
| Wang et al. (2017a) | Breast | EMT-6,1 × 106 | BALB/c, NR, 4–5 w | 11,11 | APS + DOX | 100 mg/kg, p.o., q3d ×7 | 20 mg/kg, i.v., q3d ×7 | When tumor volume reached 80 mm3 | 22 | ②⑥⑩⑪ |
| Zhu et al. (2017) | Glioma | C6, 1 × 104 | SD, NR, 6 m,200–220g | 15,15 | APS + TMZ | 250 mg/kg, i.p., qd ×15 | 20 mg/kg, p.o., qd ×5 | Post-modeling | 16 | ②⑤⑮ |
| Zhou et al. (2017b) | NPC | CNE-2,1 × 106 | BALB/c-Nu,M, 4-5w,18–20g | 5,5 | APS + DDP | 40 mg/kg,i.p., biw ×8 | 5 mg/kg, i.p., biw ×8 | After successful model establishment | 28 | ①②⑫ |
| Li et al. (2018) | Lung | LLC, 2 × 106 | C57BL/6J,F/M 6–8 w,18–20g | 10,10 | APS + DDP | 200 mg/kg, p.o., qd ×21 | 6 mg/kg, i.p., qw ×3 | 1 days post-modeling | 21 | ①②⑤⑧⑨⑮ |
| Phacharapiyangkul et al. (2019) | Melanoma | B16F10, NR | C57BL/6, NR | 7,7 | APS + DDP | 50 mg/kg, i.p., q3d ×3 | 2 mg/kg, i.p., q3d ×3 | 7 days post-modeling | 15 | ② |
| Lung | LLC, NR | C57BL/6, NR | 7,7 | APS + DDP | 50 mg/kg, i.p., q3d ×3 | 2 mg/kg, i.p., q3d ×3 | 7 days post-modeling | 15 | ② | |
| Bamodu et al. (2019) | Lung | LLC, 1.5 × 103 | C57BL/6, F, 6-8w,22.7–26.3g | 10,10 | APS + DDP | 3 mg/kg, i.p., q2w ×8 | 0.5 mg/kg, i.p., qd ×112 | When tumor volume reached 200 mm3 | 112 | ①②⑦ |
| Pan (2020) | Liver | HepG2,2 × 108 | BALB/c-Nu, F/M, 6–8 w | 7,7 | APS + DDP + DOX | 200 mg/kg, i.v., qd ×10 | 2 mg/kg DDP+ 5 mg/kg DOX, i.v., qd ×10 | 10 days post-modeling | 14 | ①②⑧⑨ |
| Pang (2020) | Breast | 4T1, 5 × 105 | BALB/c, F | 5,5 | APS-AuNP + PTX | 30 mg/kg, i.v., q3d ×6 | 6 mg/kg, i.v., q3d ×6 | When tumor volume reached 70–80 mm3 | 21 | ①②⑥⑦ |
| Li et al. (2020) | Breast | 4T1, 1.5 × 106 | BALB/c, F, 6–8 w | 15,15 | APS+5-FU | 200 mg/kg, i.p., qd ×14 | 20 mg/kg, i.p., qd ×14 | When tumor volume reached 100 mm3 | 15 | ①②④⑤⑫ |
| Wang (2021) | Colon | CT-26,2 × 106 | BALB/c, M, 4–6 w,16–20g | 5,5 | APS+5-FU | 200 mg/mL, i.p., q.d.×7 | 10 mg/mL, i.p., qd ×7 | 7 days post-modeling | 7 | ①⑮ |
| Zhang et al. (2022c) | Lung | A548/DDP, NR | BALB/c-Nu, M 6 w,18–22g | 5,5 | APS + DDP | 300 mg/kg.p.o.,qd ×24 | 3.5 mg/kg, i.p., biw ×6 | 8 days post-modeling | 32 | ①⑦⑭⑮ |
| Qu (2022) | Lung | LLC, 2 × 106 | KM mice, F, 6–8 w,18–22g | 10,10 | APS + DDP | 100 mg/kg, i.p., qd ×14 | 20 mg/kg, i.p., qd ×3 | 7 days post-modeling | 14 | ①④⑤⑮ |
| Gong et al. (2022) | Melanoma | B16/DDP, 1 × 106 | C57BL/6, M, 4–5 w | 6,6 | APS + DDP | 200 mg/kg, i.v., qd ×7 | 2.5 mg/kg, i.p., q3d ×2 | 1 day post-modeling | 14 | ①②⑬⑮ |
| Wang (2023) | Colon | CT-26,1 × 106 | BALB/c, F, 4–6 w,16–20g | 10,10 | APS+5-FU | 200 mg/kg, i.p., qd ×7 | 20 mg/kg, i.p., qd ×7 | 7 days post-modeling | 14 | ②④⑥⑬ |
| Li et al. (2023) | Liver | Hep3B,2 × 106 | BALB/c-Nu,M, 4–6 w | 5,5 | APS + DOX | 50 mg/kg, i.p., q3d ×7 | 2 mg/kg, i.p., q3d ×7 | 7 days post-modeling | 28 | ①②⑭ |
| Chen (2024) | Bladder | UM-UC-3, 1 × 107 | BALB/c-Nu, F, 4 w | 4,4 | APS + DDP | 100 mg/kg, i.p., qd ×21 | 2 mg/kg, i.p., qw ×4 | 7 days post-modeling | 28 | ① |
| Qiu et al. (2024) | Pancreas | PANC-1, NR | SD-Nude, M, 140–160g | 10,10 | APS + GEM | 200 ul/mouse, p.o., qd ×21 | 100 mg/kg, i.p., qw ×2 | After successful model establishment | 21 | ①②⑫ |
| Ma (2024) | Lung | LLC, 2 × 106 | BALB/c, F, 4–5 w,16–20g | 10,10 | APS + DDP | 200 mg/kg, p.o., qd ×21 | 2 mg/kg, i.p., qod ×7 | Within 24 h post-modeling | 21 | ①④⑤ |
| Sun et al. (2025) | Breast | 4T1, 1 × 104 | BALB/c, F, 6 w,16–18g | 8,8 | APS + DDP | 100 mg/kg, i.p., qod ×5 | 3 mg/kg, i.p., q4d ×3 | 8 days post-modeling | 18 | ②⑥⑫ |
| Chen et al. (2025) | Cervical | Hela, 1 × 106 | BALB/c-Nu, F | 7,7 | APS + DDP | 200 mg/kg, i.p., qd ×14 | 5 mg/kg, i.p., biw ×4 | After successful model establishment | 14 | ①②⑫⑮ |
| Shi et al. (2025) | Lung | LLC, 1 × 106 | BALB/c, M, 5 w,18–20g | 15,15 | APS + DDP | 400 mg/kg, i.p., qd ×14 | 6 mg/kg, i.p., qw ×2 | 24 h post-modeling | 14 | ④⑤⑦ |
Characteristics of studies included in the meta-analysis.
NPC, nasopharyngeal carcinoma; NR, no report; iv., intravenous injection; i. p., intraperitoneal injection; p. o., oral gavage; s. c., subcutaneous injection; M, male; F, female; w, weeks; qd, once daily; qod, every other day; q3d, every 3 days; q4d, every 4 days; qw, once weekly; biw, twice a week; q2w, every 2 weeks; APS, astragalus polysaccharide; RT, radiotherapy; DDP, cisplatin; TMZ, temozolomide; DOX, doxorubicin; PTX, paclitaxel; 5-FU, 5-Fluorouracil; GEM, gemcitabine; ① Tumor weight; ② Tumor volume; ③ Survival time; ④ Immune organ index; ⑤ Immune cytokines; ⑥ Immune cells; ⑦ Lung metastatic nodules; ⑧ Peripheral blood cells; ⑨ Bone marrow cells; ⑩ Liver function; ⑪ Kidney function; ⑫ Apoptosis-related molecules; ⑬ Immune checkpoint molecules; ⑭ Drug resistance-related molecules; ⑮ Others.
A total of 748 animals were evaluated across eight rodent models, including BALB/c mice (n = 11 studies), C57BL/6 mice (n = 11), KM mice (n = 6), BALB/c nude mice (n = 6), C57BL/6J mice (n = 4), C57BL/6N mice (n = 2), NOD/SCID mice (n = 2), and Sprague-Dawley (SD) rats (n = 2). Group sample sizes ranged from 4 to 15 animals. Male-only models constituted 47% (n = 21) of studies, female-only 30% (n = 13), and mixed-sex models 14% (n = 6), while 9% (n = 4) did not report sex.
APS was primarily investigated as an adjunct to chemotherapy (n = 28, 63.6%), immunotherapy (n = 12, 27.3%; including immune checkpoint inhibitors [n = 5], adoptive cell therapy [n = 4], and cancer vaccines [n = 3]), and radiotherapy (n = 4, 9.1%). Administration routes included intraperitoneal injection (i.p., n = 25, 56.8%), oral gavage (p.o., n = 12, 27.3%), intravenous infusion (i.v., n = 6, 13.6%), and subcutaneous injection (s.c., n = 1, 2.3%). APS sources were categorized into commercial research-grade preparations (n = 28, 63.6%), clinically formulated injections (n = 10, 22.7%), and laboratory-extracted preparations (n = 6, 13.6%). APS dosages ranged from 3 to 400 mg/kg/day, administered over 3–24 days.
3.3 Quality assessment of the included studies
The risk of bias across all included studies was evaluated using the SYRCLE Risk of Bias tool, as summarized in Figure 2; Supplementary Table S1. For selection bias, random sequence generation was rated low risk in 4 studies that explicitly used random number tables and unclear risk in 7 studies that only implied randomization, while 33 studies merely reported “random allocation” without providing methodological details. Allocation concealment was judged unclear risk in all 44 studies, as none described concealment procedures. Baseline characteristics were considered low risk in 42 studies that reported comparable groups, whereas 2 studies were rated unclear risk due to insufficient baseline information. For performance bias, 28 studies were judged low risk for reporting standardized housing conditions, while 16 studies were rated unclear risk because housing details were inadequately documented. Detection bias was rated low risk in only 2 studies that employed randomized outcome assessment, whereas 42 studies were considered unclear risk owing to unreported assessment procedures. Moreover, none of the 44 studies reported blinding of investigators or outcome assessors. Attrition bias was deemed low risk in all studies, as outcome data were complete. Reporting bias was also judged low risk, with no evidence of selective outcome reporting. No additional sources of bias were identified following a systematic assessment of potential confounding factors.
FIGURE 2
3.4 Outcomes and effectiveness
3.4.1 Tumor weight
Tumor weight, a primary endpoint in preclinical oncology models, may provide a direct and quantitative indicator of antitumor efficacy. Meta-analysis of 30 publications (32 independent studies) demonstrated that APS combination therapy was associated with a significant reduction in tumor weight compared with conventional monotherapy (SMD = −2.38, 95% CI [-2.95, −1.82], p < 0.00001; Figure 3). Given substantial heterogeneity (I2 = 75%; p < 0.001), we conducted meta-regression and subgroup analyses to explore potential sources. Under the random-effects model, meta-regression identified no significant associations between heterogeneity and any prespecified covariates—including species, model type, cancer type, treatment regimen, APS source, treatment initiation timing, treatment duration, or administration route (all p ≥ 0.05; Supplementary Table S2).
FIGURE 3
However, subgroup analyses identified model type (immunocompetent vs. immunodeficient; I2 = 70.2%, p = 0.07), administration route (I2 = 80.7%, p = 0.006), and treatment initiation timing (I2 = 60.2%, p = 0.04) as significant contributors to heterogeneity (Table 2). In contrast with the overall results, subgroup analyses by cancer type showed no statistically significant reduction in tumor weight in breast cancer models (SMD = −1.90, 95% CI [-3.91, 0.10], p = 0.06) or colon cancer models (SMD = −2.13, 95% CI [-5.69, 1.43], p = 0.24), indicating that the therapeutic advantage of APS may not evident in these two cancer types. Notably, the reduction in tumor weight was greater when APS was combined with immunotherapy (SMD = −2.71, 95% CI [-3.89, −1.53], p < 0.00001) compared to chemotherapy (SMD = −2.35, 95% CI [-3.04, −1.66], p < 0.00001).
TABLE 2
| Subgroup | Participants | Test for heterogeneity | Test for effect | Test for subgroup difference | |||||
|---|---|---|---|---|---|---|---|---|---|
| Studies | Com | Mon | I2 (%) | P-value | SMD [95% CI] | P-value | I2 (%) | P-value | |
| Species | |||||||||
| BALB/c | 13 | 80 | 79 | 60 | 0.0003 | −2.45 [-3.23, −1.67] | <0.0001 | 3.0 | 0.38 |
| C57BL/6 | 12 | 87 | 87 | 82 | <0.00001 | −2.23 [-3.31, −1.16] | <0.0001 | ||
| KM mice | 4 | 34 | 34 | 34 | 0.21 | −1.54 [-2.27, −0.80] | <0.0001 | ||
| NOD/SCID | 2 | 10 | 10 | 14 | 0.28 | −1.80 [-3.06, −0.54] | 0.005 | ||
| SD | 1 | 10 | 10 | - | - | - | - | ||
| Model type | |||||||||
| Immunocompetent | 23 | 168 | 167 | 74 | <0.00001 | −2.03 [-2.64, −1.42] | <0.00001 | 70.2 | 0.07 |
| Immunodeficient | 9 | 53 | 53 | 70 | 0.0007 | −3.34 [-4.60, −2.08] | <0.00001 | ||
| Tumor type | |||||||||
| Lung cancer | 11 | 87 | 87 | 81 | <0.00001 | −2.19 [-3.22, −1.17] | <0.0001 | 0 | 0.73 |
| Liver cancer | 8 | 52 | 52 | 66 | 0.005 | −2.01 [-2.99, −1.04] | <0.0001 | ||
| Melanoma | 3 | 17 | 17 | 51 | 0.13 | −2.31 [-3.76, −0.87] | 0.002 | ||
| Breast cancer | 2 | 11 | 10 | 62 | 0.11 | −1.90 [-3.91, 0.10] | 0.06 | ||
| Cervical cancer | 2 | 17 | 17 | 68 | 0.08 | −3.44 [-5.77, −1.11] | 0.004 | ||
| Colon cancer | 2 | 8 | 8 | 76 | 0.04 | −2.13 [-5.69, 1.43] | 0.24 | ||
| Others | 4 | 29 | 29 | 87 | <0.0001 | −4.71 [-7.95, −1.47] | 0.004 | ||
| Combined therapy | |||||||||
| Immunotherapy | 10 | 62 | 62 | 75 | <0.00001 | −2.71 [-3.89, −1.53] | <0.00001 | 0 | 0.61 |
| Chemotherapy | 21 | 155 | 154 | 77 | <0.00001 | −2.35 [-3.04, −1.66] | <0.00001 | ||
| Radiotherapy | 1 | 4 | 4 | - | - | - | - | ||
| APS Source | |||||||||
| Clinical-grade Injection | 6 | 48 | 47 | 56 | 0.05 | −1.73 [-2.52, −0.95] | <0.0001 | 31.6 | 0.23 |
| Research grade | 21 | 141 | 141 | 79 | <0.00001 | −2.71 [-3.54, −1.88] | <0.00001 | ||
| Self-extracted | 5 | 32 | 32 | 73 | 0.005 | −2.46 [-3.97, −0.94] | 0.001 | ||
| Administration route | |||||||||
| i.p | 16 | 115 | 115 | 50 | 0.01 | −1.63 [-2.11, −1.14] | <0.00001 | 80.7 | 0.006 |
| i.v | 6 | 39 | 39 | 41 | 0.13 | −3.08 [-4.10, −2.06] | <0.00001 | ||
| p.o | 9 | 63 | 63 | 88 | <0.00001 | −3.93 [-5.86, −2.00] | <0.0001 | ||
| s.c | 1 | 4 | 4 | - | - | - | - | ||
| Treatment initiation | |||||||||
| 1–6 days post-modeling | 13 | 92 | 92 | 76 | <0.00001 | −2.54 [-3.49, −1.59] | <0.00001 | 64.2 | 0.04 |
| 7–12 days post-modeling | 10 | 64 | 64 | 45 | 0.06 | −1.77 [-2.41, −1.14] | <0.00001 | ||
| When tumor volume reached 50–200 mm3 | 6 | 43 | 42 | 72 | 0.003 | −1.65 [-2.80, −0.51] | 0.005 | ||
| NR | 3 | 22 | 22 | 75 | <0.00001 | −6.76 [-10.47,-3.05] | 0.0004 | ||
| Duration | |||||||||
| ≤14 days | 12 | 85 | 85 | 70 | 0.0001 | −2.04 [-2.84, −1.25] | <0.00001 | 11.4 | 0.29 |
| >14 days | 20 | 136 | 135 | 78 | <0.00001 | −2.66 [-3.46, −1.85] | <0.00001 | ||
Subgroup analysis for tumor weight.
3.4.2 Tumor volume
Tumor volume is commonly used as another primary endpoint for assessing antitumor efficacy. Analysis of 24 articles (27 studies) revealed that combined APS and antitumor therapy was associated with a significant reduction in tumor volume compared to monotherapy (SMD = −2.70, 95% CI [-3.36, −2.04], p < 0.00001; Figure 4). Substantial heterogeneity was observed across the studies (I2 = 79%, p < 0.00001). Meta-regression of prespecified covariates identified no significant heterogeneity sources (all p ≥ 0.05; Supplementary Table S2). Subgroup analyses suggested that the combination regimen (I2 = 60.1%, p = 0.08) and treatment duration (I2 = 84.4%, p = 0.01) were significant contributors to heterogeneity. Similarly, in breast cancer models (SMD = −3.14, 95% CI [-7.91, 1.62], p = 0.20) and colon cancer models (SMD = −1.74, 95% CI [-5.94, 2.46], p = 0.42), APS combination therapy did not produce a statistically significant reduction in tumor volume compared with monotherapy controls (Table 3).
FIGURE 4
TABLE 3
| Subgroup | Participants | Test for heterogeneity | Test for effect | Test for subgroup difference | |||||
|---|---|---|---|---|---|---|---|---|---|
| Studies | Com | Mon | I2 (%) | P-value | SMD [95% CI] | P-value | I2 (%) | P-value | |
| Species | |||||||||
| BALB/c | 8 | 51 | 51 | 83 | <0.00001 | −3.03 [-4.60, −1.46] | 0.0002 | 0 | 0.65 |
| C57BL/6 | 13 | 91 | 91 | 79 | <0.0001 | −3.03 [-4.07, −1.98] | <0.00001 | ||
| KM mice | 2 | 20 | 20 | 57 | 0.13 | −2.28 [-3.59, −0.98] | 0.006 | ||
| NOD/SCID | 2 | 10 | 10 | 0 | 0.72 | −1.91 [-3.07, −0.75] | 0.001 | ||
| SD rat | 2 | 25 | 25 | 93 | 0.0001 | −2.46 [-6.10, 1.19] | 0.19 | ||
| Model type | |||||||||
| Immunocompetent | 20 | 153 | 152 | 81 | <0.0001 | −2.62 [-3.40, −1.83] | <0.00001 | 0 | 0.68 |
| Immunodeficient | 7 | 44 | 44 | 42 | 0.11 | −2.87 [-3.79, −1.96] | <0.00001 | ||
| Tumor type | |||||||||
| Lung cancer | 7 | 56 | 56 | 82 | <0.0001 | −2.69 [-4.04, −1.34] | <0.0001 | 0 | 0.52 |
| Liver cancer | 7 | 45 | 45 | 65 | 0.009 | −3.51 [-4.82, −2.19] | <0.0001 | ||
| Melanoma | 4 | 24 | 24 | 78 | 0.004 | −2.71 [-4.60, −0.83] | 0.005 | ||
| Breast cancer | 2 | 11 | 10 | 85 | 0.01 | −3.14 [-7.91, 1.62] | 0.20 | ||
| Cervical cancer | 2 | 17 | 17 | 0 | 0.65 | −1.83 [-2.67, −0.99] | <0.0001 | ||
| Colon cancer | 2 | 14 | 14 | 85 | 0.01 | −1.74 [-5.94, 2.46] | 0.42 | ||
| Others | 3 | 30 | 30 | 90 | <0.00001 | −3.12 [-6.19, −0.06] | 0.05 | ||
| Combined therapy | |||||||||
| Immunotherapy | 10 | 62 | 62 | 53 | 0.02 | −2.10 [-2.85, −1.35] | <0.00001 | 60.1 | 0.08 |
| Radiotherapy | 3 | 30 | 30 | 61 | 0.08 | −1.84 [-2.87, −0.81] | 0.0005 | ||
| Chemotherapy | 14 | 105 | 104 | 87 | <0.00001 | −3.59 [-4.85, −2.32] | <0.00001 | ||
| APS Source | |||||||||
| Clinical-grade Injection | 8 | 67 | 66 | 79 | <0.0001 | −2.64 [-3.78, −1.50] | <0.00001 | 0 | 0.97 |
| Research grade | 15 | 108 | 108 | 82 | <0.00001 | −2.83 [-3.82, −1.85] | <0.00001 | ||
| Self-extracted | 4 | 22 | 22 | 75 | 0.008 | −2.74 [-4.67, −0.81] | 0.005 | ||
| Administration route | |||||||||
| i.p | 13 | 102 | 101 | 79 | <0.00001 | −2.36 [-3.24, −1.49] | <0.00001 | 0 | 0.59 |
| i.v | 5 | 29 | 29 | 70 | 0.01 | −3.03 [-4.64, −1.41] | 0.0002 | ||
| p.o | 8 | 62 | 62 | 85 | <0.00001 | −3.14 [-4.61, −1.68] | <0.0001 | ||
| s.c | 1 | 4 | 4 | - | - | - | - | ||
| Treatment initiation | |||||||||
| 1–6 days post-modeling | 9 | 61 | 61 | 82 | <0.00001 | −2.78 [-4.14, −1.43] | <0.0001 | 0 | 0.72 |
| 7–12 days post-modeling | 9 | 71 | 71 | 84 | <0.00001 | −2.76 [-4.00, −1.52] | <0.0001 | ||
| When tumor volume reached 50–200 mm3 | 6 | 43 | 42 | 63 | 0.02 | −2.31 [-3.36, −1.25] | <0.0001 | ||
| NR | 3 | 22 | 22 | 63 | 0.07 | −3.54 [-5.38, −1.70] | 0.002 | ||
| Duration | |||||||||
| ≤14 days | 12 | 97 | 97 | 69 | 0.0002 | −1.90 [-2.59, −1.22] | <0.00001 | 84.4 | 0.01 |
| >14 days | 15 | 100 | 99 | 83 | <0.00001 | −3.72 [-4.94, −2.49] | <0.00001 | ||
Subgroup analysis for tumor volume.
3.4.3 Survival and metastasis outcomes
Survival time is the most clinically relevant endpoint for evaluating antitumor efficacy. Notably, only 2 of the 44 included studies (4.5%) reported analyzable survival data, substantially limiting the generalizability of these findings. Pooled analysis of these limited studies showed a significant prolongation of median survival time with APS combination therapy compared to monotherapy (MD = 12.92 days, 95% CI [10.88, 14.96], p < 0.00001; Figure 5A). Pulmonary metastatic nodule count, an established quantitative measure of metastatic burden and an independent prognostic factor, was assessed as a secondary efficacy endpoint. Pooled data from four studies demonstrated that APS-chemotherapy regimens were associated with a significant reduction in lung metastasis (SMD = −3.67, 95% CI [-5.74, −1.60], p = 0.0005; Figure 5B).
FIGURE 5
3.4.4 Immune organ index
The thymus and spleen indices—calculated as organ weight-to-body weight ratios—could serve as quantitative indicators of immunoenhancement or immunosuppression during antitumor therapy. Pooled analyses showed that APS combination therapy significantly increased the spleen index (10 studies; SMD = 1.70, 95% CI [0.79, 2.61], p = 0.0003) and thymus index (8 studies; SMD = 1.82, 95% CI [0.76, 2.88], p = 0.0007) compared with monotherapy (Figure 6).
FIGURE 6
3.4.5 CD4+ T cells and CD8+ T cells
Increased CD8+ T-cell infiltration in the tumor, spleen, and draining lymph nodes generally reflects enhanced systemic antitumor immune activation. Meta-analysis of 14 studies showed that APS combination therapy significantly increased CD8+ T-cell proportions in tumor tissue (SMD = 3.17, 95% CI [1.00, 5.33], p = 0.004), spleen (SMD = 1.84, 95% CI [0.75, 2.94], p = 0.001) and draining lymph nodes (SMD = 1.36, 95% CI [0.23, 2.49], p = 0.02), respectively. Notably, peripheral blood CD8+ T-cell frequencies remained unchanged (SMD = 0.72, 95% CI [−2.90, 4.34], p = 0.70). In contrast, pooled analyses of 10 studies revealed that APS combination therapy significantly elevated peripheral blood CD4+ T-cell frequencies (SMD = 1.92, 95% CI [0.84, 3.00], p = 0.0005), while tumor-infiltrating and lymphoid tissue-resident CD4+ T-cell populations showed no significant alterations (all p > 0.05; Figure 7). This compartmentalized distribution pattern may reflect distinct trafficking dynamics between CD4+ and CD8+ T-cell subsets in response to APS-enhanced immunotherapy.
FIGURE 7
3.4.6 Cytokine levels
As pivotal immune regulators, cytokines orchestrate antitumor responses both within the tumor microenvironment and systemically. In the included preclinical studies, APS combination therapy was reported to alter several serum cytokines, although interpretation is constrained by notable methodological limitations and substantial between-study heterogeneity. Pooled analyses indicated higher concentrations of tumor necrosis factor-α (TNF-α: SMD = 8.96, 95% CI [4.96, 12.95], p < 0.0001), interleukin-2 (IL-2: SMD = 4.54, 95% CI [1.57, 7.51], p = 0.003), interferon-γ (IFN-γ: SMD = 4.29, 95% CI [1.80, 6.78], p = 0.0007), interleukin-12 (IL-12: SMD = 3.66, 95% CI [2.53, 4.80], p < 0.00001), and interleukin-6 (IL-6: SMD = 11.07, 95% CI [1.44, 20.70], p = 0.02) in the APS combination groups. A reduction in interleukin-10 was also observed (IL-10: SMD = −8.47, 95% CI [-16.71, −0.23], p = 0.04). No statistically significant differences were detected for transforming growth factor-β (TGF-β: SMD = −1.96, 95% CI [-5.16, 1.24], p = 0.23) or interleukin-1β (IL-1β: SMD = 2.22, 95% CI [-3.62, 8.07], p = 0.46). (Figures 8, 9).
FIGURE 8
FIGURE 9
3.4.7 Immune checkpoint
Programmed cell death protein 1 (PD-1) and its ligand PD-L1 represent critical immune checkpoint molecules that mediate tumor immune evasion. Among the included studies, three provided evaluable PD-1/PD-L1 expression data. Meta-analysis revealed that APS combination therapy significantly downregulated PD-1/PD-L1 expression in tumor tissues (SMD = −3.66, 95% CI [-6.11, −1.20], p = 0.003; Figure 10).
FIGURE 10
3.4.8 Safety assessment
The safety profile of APS combination therapy was evaluated in eight included studies, which assessed general animal condition, hepatic and renal function parameters, and organ-specific histopathology. Of these, four studies (Huang et al., 2008; Tian, 2012; Ma, 2024; Chen et al., 2025) assessing general health status—such as daily activity, mental state, and coat appearance—reported no APS-related adverse effects. Notably, two of these studies indicated that mice in the combination therapy groups exhibited better mental status compared to those receiving chemotherapy alone. Histological evaluations of organs in three studies (Wang et al., 2017b; Ma, 2024; Sun et al., 2025) showed no signs of drug-induced tissue injury.
Based on data availability, serum creatinine (CRE), blood urea nitrogen (BUN), aspartate aminotransferase (AST), and alanine aminotransferase (ALT) were incorporated into the meta-analysis. The result showed that APS combination therapy was associated with improved renal function, evidenced by decreased CRE levels (SMD = −1.15, 95% CI [-1.82, −0.48], p = 0.0008). A trend toward hepatoprotection was also observed, with ALT levels approaching statistical significance (SMD = −1.65, 95% CI [-3.36, 0.06], p = 0.06), whereas AST and BUN levels did not change significantly (Figure 11).
FIGURE 11
In addition, two studies evaluated hematological outcomes. Both studies consistently reported that APS combined with chemotherapy led to higher counts of bone marrow cells, platelets (PLT), white blood cells (WBC), and red blood cells (RBC) relative to chemotherapy alone, suggesting a potential hematopoietic-supportive effect of APS.
3.5 Sensitivity analysis
Sensitivity analysis confirmed the robustness of the primary outcomes—including tumor weight and volume, survival time, metastatic nodules, immune organ indices (thymus and spleen), T-cell subset frequencies (CD4+/CD8+), and key cytokines (IL-12, IFN-γ, TNF-α, IL-10, and IL-2)—as the iterative exclusion of individual studies did not materially affect pooled effect sizes or heterogeneity patterns. Four outcomes exhibited sensitivity to individual study exclusion. Specifically, PD-1/PD-L1 effects lost statistical significance (p > 0.05) upon removal of Gong et al. (2022); IL-6 lost significance (p > 0.05) when any of three studies were omitted (Ming et al., 2014; Tian, 2012; Shi et al., 2025); IL-10 similarly lost significance when excluding Zhu et al. (2017), Tian (2012), or Li et al. (2018); conversely, IL-1β became statistically significant (p < 0.05) after exclusion of Zhang (2019).
3.6 Publication bias and evidence quality
Publication bias assessment for outcomes with ten or more studies revealed no evidence of bias for CD4+ T-cell frequencies or the spleen index, as indicated by symmetrical funnel plots and non-significant Egger’s tests (p > 0.05). In contrast, significant funnel asymmetry, supported by Egger’s test (p < 0.05), suggested potential publication bias for tumor volume, tumor weight, and CD8+ T-cell frequencies (Figures 12, 13).
FIGURE 12
FIGURE 13
Trim-and-fill analyses indicated that two imputed studies modestly reduced the effect size for tumor weight (adjusted SMD = −0.52, 95% CI [-0.64, −0.40] vs. original SMD = −0.48, 95% CI [-0.60, −0.36], ΔSMD = −0.04), although statistical significance was retained (p < 0.001). In contrast, no imputed studies were generated for tumor volume or CD8+ T-cell frequencies, and their effect sizes remained unchanged (Supplementary Table S3). According to the GRADE framework, the certainty of evidence for the 21 evaluated outcomes ranged from very low to low. Low-certainty ratings were assigned to six outcomes—thymus index, lung metastatic nodule count, survival time, TNF-α, IFN-γ, and IL-12—while the remaining fifteen outcomes were rated as very low certainty. Detailed GRADE assessments and downgrading domains—including risk of bias, inconsistency, indirectness, imprecision, and publication bias—are provided in Supplementary Table S4.
4 Discussion
4.1 Summary of the evidence and main findings
Many natural polysaccharides are increasingly recognized as valuable adjuvants in cancer combination therapy (Li et al., 2021; Ornelas et al., 2022). In particular, APS, a bioactive extract from Astragali Radix, has attracted considerable attention and is being extensively investigated for its role in potentiating conventional antitumor treatments (Bamodu et al., 2019; Xu et al., 2023; He et al., 2024). Therefore, we herein report the first systematic review and meta-analysis to comprehensively evaluate the efficacy, safety, and mechanistic basis of APS as an adjunctive therapy in animal tumor models.
Our synthesis of 44 preclinical studies (n = 748 animals) indicates that APS co-administration significantly reduced tumor burden, suppressed pulmonary metastasis, and moderately improved survival (all p < 0.05), potentially by enhancing T-cell immunity and upregulating antitumor cytokines. A favorable safety profile was also observed. However, according to GRADE assessments, the certainty of evidence for all primary outcomes was low or very low, indicating that these findings should be interpreted cautiously and require validation in rigorously controlled preclinical studies.
4.2 Exploration of heterogeneity and its Implications
Substantial heterogeneity was observed across most primary efficacy outcomes. Meta-regression analyses of eight covariates (e.g., species, model, therapy type) failed to identify significant heterogeneity sources (all p ≥ 0.05; Supplementary Table S3). This implies that the heterogeneity likely stems from multiple interacting methodological and biological factors rather than a single dominant contributor.
Subgroup analyses identified administration route as a major contributor to heterogeneity in tumor weight (between-subgroup I2 = 80.7%, p = 0.006), with oral APS demonstrating superior efficacy compared to parenteral delivery. This pattern may reflect route-dependent differences in pharmacokinetic and immunological engagement. Specifically, oral APS undergoes M cell-mediated transcytosis in Peyer’s patches, activating dendritic cell TLR4 signaling and initiating gut-associated lymphoid tissue (GALT)-primed systemic immunity (Zhang et al., 2022a; 2022b); unabsorbed polysaccharides are subsequently metabolized by colonic microbiota into short-chain fatty acids, further enhancing immunomodulation (Ye et al., 2023). In contrast, parenteral administration bypasses enteric immune activation and relies primarily on direct interaction between circulating APS and systemic immune cells, making its efficacy more dependent on precise structural properties (Cai et al., 2023; Wang et al., 2024). While compelling, these mechanistic explanations remain preliminary due to insufficient reporting of phytochemical and PK/PD data, rendering any definitive conclusion premature. Thus, future studies employing well-characterized APS preparations and detailed PK/PD analyses are essential for validation.
A critical yet unquantified source of between-study variability is the structural heterogeneity of APS itself. The included studies used APS derived from various sources, each subject to distinct extraction and purification procedures that influence molecular weight distribution, branching patterns, and receptor-binding properties (Tang and Huang, 2022). Although subgroup analyses did not detect statistically significant differences among APS sources, this null finding may reflect low statistical power or coarse categorization rather than true equivalence.
Model immune status was identified as a notable contributor to heterogeneity. Subgroup analysis revealed numerically stronger synergistic effects in immunodeficient models (tumor weight: SMD = −3.34; tumor volume: SMD = −2.87) compared to immunocompetent models (tumor weight: SMD = −2.03; tumor volume: SMD = −2.62). This pattern should not be interpreted as superior APS efficacy in the absence of immunity, but may reflect the predominance of immune-independent mechanisms (e.g., direct apoptosis or autophagy induction) combined with reduced biological variability inherent to immunodeficient systems. However, this finding derives from subgroup analysis with limited mechanistic data, warranting cautious interpretation and rigorous experimental validation.
Additional sources of heterogeneity include differences in treatment timing, intervention duration, and the combination therapy. Tumor type-specific responses, such as no significant effects observed in breast and colon cancer models, likely reflect tumor microenvironment responsiveness to APS or low statistical power from an insufficient number of studies. Unmeasured experimental factors—such as tumor inoculation protocols, cell line characteristics, housing conditions, and outcome assessment methods—may further contribute to cross-study inconsistencies.
4.3 Potential mechanisms underlying the observed effects
Building upon the subgroup findings, mechanistic indicators extracted from the included studies (Table 4) suggest that APS may influence antitumor responses through both immune-dependent and immune-independent pathways. Immune-mediated mechanisms appear to contribute substantially to APS activity. Increases in spleen and thymus indices in tumor-bearing mice imply that APS co-administration may mitigate lymphoid organ atrophy associated with tumor progression or chemotherapy, thereby preserving the structural basis for immune activation. Enhancements in CD8+ T-cell infiltration within tumors and secondary lymphoid organs, together with elevated levels of cytokines such as IFN-γ, TNF-α, IL-2, and IL-12, further indicate partial restoration of antitumor immunity. Concurrent reductions in PD-1/PD-L1 expression and IL-10 levels suggest attenuation of immunosuppressive signaling. Although the mechanistic signals identified in our analysis are constrained by low-quality evidence, similar immune-enhancing effects have been reported in clinical studies of APS injection, including improvements in immune cell subsets (Li and Jiang, 2024) and modulation of inflammatory cytokines (Huang et al., 2019; Mok et al., 2024). This convergence between preclinical and clinical findings provides preliminary support for APS-mediated immunomodulation, although the underlying mechanisms will require further confirmation in rigorously controlled studies.
TABLE 4
| Outcome measures | Counts | Outcome measures | Counts |
|---|---|---|---|
| Tumor weight | 32 | TGF-β levels of blood | 3 |
| Tumor volume | 27 | IL-10 levels of tumor | 3 |
| T cells of tumor | 11 | TGF-β levels of tumor | 3 |
| Spleen index | 10 | TNF-α levels of tumor | 3 |
| Thymus index | 8 | T cells of Lymph node | 3 |
| TNF-α levels of blood | 7 | VEGF | 3 |
| IL-2 levels of blood | 7 | p53 (Western blot) | 2 |
| IFN-γ levels of blood | 6 | p53 mRNA | 2 |
| Lung metastatic nodules | 4 | P-glycoprotein (Western blot) | 2 |
| T cells of spleen | 4 | Liver index | 2 |
| T cells of blood | 4 | Caspase-3 (Western blot) | 2 |
| IL-10 levels of blood | 4 | Bone marrow cells | 2 |
| IL-1β levels of blood | 4 | Platelets (PLT) | 2 |
| IL-6 levels of blood | 4 | PLT (RBC) | 2 |
| Alanine aminotransferase (ALT) | 4 | White blood cells (WBC) | 2 |
| Aspartate aminotransferase (AST) | 3 | IL-12 levels of blood | 2 |
| Creatinine (CRE) | 3 | Survival time (days) | 2 |
| Urea nitrogen (UBN) | 3 | Bax (Western blot) | 2 |
| PD-1/PD-L1 of tumor | 3 | Bcl-2 (Western blot) | 2 |
Outcome measures reported ≥2 times.
Notably, eight included studies (nine experiments) employed immunodeficient mouse models. Because mechanistic data were limited in these experiments, only tumor-related endpoints were analyzed. Accordingly, all immune-related analyses in this meta-analysis were conducted exclusively in immunocompetent models. APS exhibited tumor-inhibitory effects in both immunocompetent and immunodeficient hosts, suggesting that adaptive immunity may not be the sole contributor to its activity. In nude mouse models, APS increased the Bax/Bcl-2 ratio (Zhou Z. et al., 2017; Qiu et al., 2024), modulated autophagy-related proteins (LC3-II, Beclin-1, p62) (Qiu et al., 2024; Chen et al., 2025), and reduced the expression of drug-resistance transporters (P-gp, MRP, LRP) (Pan, 2020; Zhang Y. et al., 2022), indicating potential involvement of apoptosis induction, autophagy regulation, and enhanced intracellular drug accumulation. Nevertheless, these mechanistic insights should be interpreted cautiously, as they originate from a small number of studies, were synthesized narratively rather than meta-analytically, and often lacked comprehensive pathway-level validation. Therefore, while the dual-pathway model provides a plausible framework, the precise contribution and interplay of immune-dependent and independent mechanisms remain a critical area for future rigorous investigation.
Beyond its potential antitumor mechanisms, APS may also attenuate treatment-related toxicities. In this meta-analysis, APS administration was associated with significantly reduced serum creatinine levels, suggesting possible renal protection. In contrast, hepatic enzymes (AST and ALT) showed no significant alterations, indicating preserved hepatic function. Limited qualitative evidence suggested possible mitigation of chemotherapy-induced myelosuppression and potential hematopoietic support (Li et al., 2018; Pan, 2020). However, because only eight included studies reported safety-related outcomes, the current evidence remains insufficient to draw definitive conclusions.
Overall, APS may exert complementary effects through immune modulation, cytoprotection, and support of host physiological resilience. These observations conceptually parallel the TCM principle of Fuzheng (扶正), which emphasizes strengthening host defenses in conjunction with Quxie (祛邪) therapies—such as chemotherapy or radiotherapy—to preserve treatment tolerance and improve therapeutic responsiveness.
4.4 Limitations and robustness of evidence
Although APS co-administration showed statistically favorable trends across multiple outcomes, the overall robustness of these conclusions is fundamentally constrained by the low quality of the primary studies. All major endpoints were rated as “low” or “very low” certainty by GRADE, reflecting pervasive methodological weaknesses—including inadequate reporting of randomization, allocation concealment, and blinding—that increase the likelihood of biased effect estimates. The substantial and largely unexplained heterogeneity further reduces confidence in pooled results, as effect size variability could stem from uncontrolled differences in APS preparation, dosing, model systems, and experimental procedures. Moreover, publication bias may have inflated positive findings, and data extraction from figures may introduce measurement error. These limitations collectively indicate that the apparent benefits of APS should be interpreted as preliminary rather than definitive. Nonetheless, the generally consistent direction of effects across diverse models provides a tentative signal warranting further investigation. Strengthening the evidence base will require rigorously designed animal studies that adhere to SYRCLE and ARRIVE guidelines.
4.5 Implications for clinical translation and future research
To bridge the gap between the preclinical promise of APS and its clinical application, several critical hurdles must be addressed. First, standardized APS preparations need to be developed. This requires both defining biologically relevant chemical attributes (e.g., molecular weight distribution, glycosidic linkage patterns, and monosaccharide composition) and implementing robust quality-control frameworks to ensure batch-to-batch consistency across sourcing, cultivation, extraction, and purification. Second, dosing strategies require systematic optimization. While existing clinical trials have utilized only injectable APS, preclinical data suggest potentially superior efficacy with oral administration. Therefore, developing and rigorously evaluating oral formulations represents a promising future direction—though this necessitates establishing new quality standards and providing mechanistic justification. Third, human safety profiles require careful evaluation. Despite preliminary evidence indicating favorable tolerability and toxicity mitigation in animal models, the evidence remains limited. Future clinical trials should incorporate systematic monitoring for immune-related adverse events (irAEs) and potential drug interactions. In summary, the successful clinical translation of APS in cancer therapy will depend on coordinated advances in standardization, dosing optimization, and comprehensive safety evaluation.
5 Conclusion
This meta-analysis provides preclinical evidence that APS may serve as an adjunctive agent to enhance the efficacy of conventional cancer therapies. However, given the low certainty of current evidence, further mechanistic studies and well-designed clinical trials are urgently warranted to establish its efficacy and therapeutic role in oncology.
Statements
Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.
Author contributions
RZ: Conceptualization, Data curation, Formal Analysis, Methodology, Software, Visualization, Writing – original draft, Writing – review and editing. QY: Data curation, Formal Analysis, Investigation, Software, Visualization, Project administration, Writing – original draft. ZC: Data curation, Investigation, Resources, Software, Writing – original draft. JH: Conceptualization, Supervision, Validation, Writing – original draft, Writing – review and editing, Visualization. GZ: Conceptualization, Funding acquisition, Resources, Supervision, Writing – original draft, Writing – review and editing, Methodology, Visualization.
Funding
The author(s) declared that financial support was not received for this work and/or its publication.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fphar.2025.1672450/full#supplementary-material
References
1
AnE. K.ZhangW.KwakM.LeeP. C. W.JinJ. O. (2022). Polysaccharides from Astragalus membranaceus elicit T cell immunity by activation of human peripheral blood dendritic cells. Int. J. Biol. Macromol.223, 370–377. 10.1016/j.ijbiomac.2022.11.048
2
BamoduO. A.KuoK. T.WangC. H.HuangW.-C.WuA. T. H.TsaiJ. T.et al (2019). Astragalus polysaccharides (PG2) enhances the M1 polarization of macrophages, functional maturation of dendritic cells, and T cell-mediated anticancer immune responses in patients with lung cancer. Nutrients11 (10), 2264. 10.3390/nu11102264
3
CaiY.SiZ.JiangY.YeM.WangF.YangX.et al (2023). Structure-activity relationship of low molecular weight astragalus membranaceus polysaccharides produced by bacteroides. Carbohydr. Polym.316, 121036. 10.1016/j.carbpol.2023.121036
4
CaoQ.ZhouR.GuoS.MengK.YangX.LiuM.et al (2024). PLGA-Astragalus polysaccharide nanovaccines exert therapeutic effect in colorectal cancer. Int. J. Nanomed.19, 9437–9458. 10.2147/IJN.S479334
5
ChangC.-C.LeeY. L.YinC. H.ChangC.-C.LinY.-H. (2025). Efficacy of astragalus polysaccharides (PG2) in alleviating chemotherapy-induced fatigue in gynecologic cancer: a retrospective cohort study. Discov. Oncol.16 (1), 1322. 10.1007/s12672-025-03120-y
6
ChenR. Q. (2024). Experiment research of astragalus polysaccharide in inhibiting bladder cancer and enhance its sensitivity to Cisplatin based on network pharmacology and experimental verification. Guangdong Med. Univ. Master’s Thesis.
7
ChenH. W.LinI. H.ChenY. J.ChangK. H.WuM. H.SuW. H.et al (2012). A novel infusible botanically-derived drug, PG2, for cancer-related fatigue: a phase II double-blind, randomized placebo-controlled study. Clin. Invest. Med.35 (1), E1–E11. 10.25011/cim.v35i1.16100
8
ChenK. Y.YangW.ZhaoC.QinJ. (2025). Preliminary study on synergistic effect of Astragalus polysaccharide in chemotherapy of cervical cancer. Chin. Arch. Tradit. Chin. Med., 1–15. 10.1546./r.20250121.1116.004
9
DuY.WanH. T.HuangP.YangJ. H.HeY. (2022). A critical review of astragalus polysaccharides: from therapeutic mechanisms to pharmaceutics. Biomed. Pharmacother.147, 112654. 10.1016/j.biopha.2022.112654
10
DurazzoA.NazhandA.LucariniM.SilvaA. M.SoutoS. B.GuerraF.et al (2021). Astragalus (astragalus membranaceus bunge): botanical, geographical, and historical aspects to pharmaceutical components and beneficial role. Rend. Lincei, Sci. Fis. Nat.32 (3), 625–642. 10.1007/s12210-021-01003-2
11
GongQ.YuH.DingG.MaJ.WangY.ChengX. (2022). Suppression of stemness and enhancement of chemosensibility in the resistant melanoma were induced by astragalus polysaccharide through PD-L1 downregulation. Eur. J. Pharmacol.916, 174726. 10.1016/j.ejphar.2021.174726
12
GuoL.BaiS. P.ZhaoL.WangX. H. (2012). Astragalus polysaccharide injection integrated with vinorelbine and cisplatin for patients with advanced non-small cell lung cancer: effects on quality of life and survival. Med. Oncol.29 (3), 1656–1662. 10.1007/s12032-011-0068-9
13
HeZ.LiuX.QinS.YangQ.NaJ.XueZ.et al (2024). Anticancer mechanism of astragalus polysaccharide and its application in cancer immunotherapy. Pharmaceuticals17 (5), 636. 10.3390/ph17050636
14
HuangH. S.HuangW. T.QuZ. J.TangY. J. (2002). A laboratory trial on the combination of CD3AK and APS in the treatment of mice with melanoma B16. J. Youjiang Med. Univ. Natl. (3), 332–334. 10.3969/j.issn.1001-5817.2002.03.003
15
HuangH. S.HuangW. T.WeiP. Y.HuangY. Q. (2008). Experimental research on astragalus polysaccharides combined with cisplatin in treating H22 liver cancer. J. Li-shizhen Tradit. Chin. Med. (11), 2741–2742. 10.3969/j.issn.1008-0805.2008.11.080
16
HuangW. C.KuoK. T.BamoduO. A.LinY. K.WangC. H.LeeK. Y.et al (2019). Astragalus polysaccharide (PG2) ameliorates cancer symptom clusters, as well as improves quality of life in patients with metastatic disease, through modulation of the inflammatory cascade. Cancers11 (8), 1054. 10.3390/cancers11081054
17
HuoH. Q. (2016). In vitro and in vivo anti-tumor effects of CIK combined with astragalus polysaccharides upon cervical cancer Hela cells. Shijiazhuang, Hebei, China: Hebei Medical University Mastre’s Thesis. 10.7666/d.D843900
18
JairamV.LeeV.ParkH. S.ThomasC. R.MelnickE. R.GrossC. P.et al (2019). Treatment-related complications of systemic therapy and radiotherapy. JAMA Oncol.5 (7), 1028–1035. 10.1001/jamaoncol.2019.0086
19
LiG. Y.JiangJ. (2024). Recent efficacy and long-term survival of astragalus polysaccharide combined with gemcitabine and S-1 in pancreatic cancer. World J. Clin. Oncol.15 (11), 1404–1411. 10.5306/wjco.v15.i11.1404
20
LiY.YangL. L.BaiY. L.MingH. X. (2018). Effects of astragalus polysaccharides on growth, metastasis and cell cycle of lewis lung cancer in tumor-bearing mice with Qi and yin deficiency. Cancer Res. Prev. Treat.45 (3), 131–137. 10.3971/j.issn.1000-8578.2018.17.0220
21
LiW.HuX.WangS.JiaoZ.SunT.LiuT.et al (2020). Characterization and anti-tumor bioactivity of astragalus polysaccharides by immunomodulation. Int. J. Biol. Macromol.145, 985–997. 10.1016/j.ijbiomac.2019.09.189
22
LiY.WangX.MaX.LiuC.WuJ.SunC. (2021). Natural polysaccharides and their derivates: a promising natural adjuvant for tumor immunotherapy. Front. Pharmacol.12, 621813. 10.3389/fphar.2021.621813
23
LiC.LiuY.ZhangY.LiJ.LaiJ. (2022). Astragalus polysaccharide: a review of its immunomodulatory effect. Arch. Pharmacal Res.45 (6), 367–389. 10.1007/s12272-022-01393-3
24
LiM.DuanF.PanZ.LiuX.LuW.LiangC.et al (2023). Astragalus polysaccharide promotes doxorubicin-induced apoptosis by reducing O-GlcNAcylation in hepatocellular carcinoma. Cells12 (6), 866. 10.3390/cells12060866
25
LiN.ZhangY.HanM. M.LiuT.WuJ. J.XiongY. X.et al (2024). Self-adjuvant astragalus polysaccharide-based nanovaccines for enhanced tumor immunotherapy: a novel delivery system candidate for tumor vaccines. Sci. China Life Sci.67 (4), 680–697. 10.1007/s11427-023-2465-x
26
LinC.WuF.LianH.WangL.BiL.WangL. (2022). Antitumor effect of combination treatment of astragalus polysaccharide and PD-L1 antibody on mice with lewis lung carcinoma. TMR Pharmacol. Res.2 (1), 4. 10.53388/PR202202004
27
LinZ. C.ZhouH. R.ZhongY. Z.LuH.WangJ. (2024). Effects of astragalus polysaccharide on inhibitory ability of PD-1 inhibitor in H22 tumor-bearing mice. Chin. J. Tradit. Med. Sci. Technol.31 (2), 217–222.
28
LiuB.ZhouH.TanL.SiuK. T. H.GuanX.-Y. (2024a). Exploring treatment options in cancer: tumor treatment strategies. Signal Transduct. Target. Ther.9 (1), 175. 10.1038/s41392-024-01856-7
29
LiuH. C.JinY. T.ZhangR. X.HeY.BianY. H. (2024b). A preliminary investigation of the mechanism of the anti-lung adenocarcinoma effect of astragalus polysaccharide by modulating T cell sensitization of PD-L1 blockers. J. Mod. Oncol.32 (9), 1573–1580. 10.3969/j.issn.1672-4992.2024.09.001
30
LiuY. X.SongX. M.DanL. W.TangJ. M.JiangY.DengC.et al (2024c). Astragali radix: comprehensive review of its botany, phytochemistry, pharmacology and clinical application. Arch. Pharmacal Res.47 (3), 165–218. 10.1007/s12272-024-01489-y
31
LvX. M.XuJ. F.YanC. X.XuD. X.SunY. Y. (2017). Effect of astragalus polysaccharide combined with radiotherapy on radiosensitivity and immune function of transplanted tumor in mice. J. Yunnan Univ. Chin. Med.40 (6), 9–13. 10.19288/j.cnki.issn.1000-2723.2017.06.003
32
MaY. C. (2024). Astragalus polysaccharides inhibits the tumor growth of Lewis lung cancer mice by regulating intestinal flora and immune function. Lanzhou, Gansu, China: Lanzhou University. Master’s Thesis. 10.27204/d.cnki.glzhu.2024.002131
33
MingH. X.ChenY. W.HuY. H.DongX. L.GuJ.LiY. (2014). Effects of astragalus polysaccharide APS on cytokine and immune function impairment induced by cisplatin in mice bearing lewis lung cancer. Acta Lab. Anim. Sci. Sin.22 (5), 44–48. 10.3969/j.issn.1005-4847.2014.05.010
34
MokL. M.HuangW. C.LinH. C.ChenY. J.ChanM. L.LaiY. L.et al (2024). 1435P immunomodulatory effects and improved survival of PG2 plus preoperative chemoradiotherapy in patients with locally advanced esophageal cancer. Ann. Oncol.35, S893. 10.1016/j.annonc.2024.08.1501
35
MuJ. Y.LiY. D.ZhaoX. K.LiJ.YangA. J. (2019). Astragalus polysaccharide restores activation of NK cells in radiation therapy of tumors. Int. J. Clin. Exp. Med.12 (7), 8609–8621.
36
National Cancer Institute (2017). Types of cancer treatment. Available online at: https://www.cancer.gov/about-cancer/treatment/types (Accessed July 20, 2025).
37
OrnelasA. C.FergusonS.DePlazaM.AdekunleT.BashaR. (2022). Anti-cancer pectins and their role in colorectal cancer treatment. Onco Ther.9 (2), 43–55. 10.1615/OncoTherap.v9.i2.50
38
PageM. J.McKenzieJ. E.BossuytP. M.BoutronI.HoffmannT. C.MulrowC. D.et al (2021). The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJn71, n71. 10.1136/bmj.n71
39
PanD. S. (2020). Chemosensitizing and chemoprotective effects of Astragalus polysaccharide in hepatoma chemotherapy: mechanisms and therapeutic implications. Mudanjiang, Heilongjiang, China: Mudanjiang Medical University. Master’s Thesis. 10.27877/d.cnki.gmdjm.2020.000002
40
PangG. B. (2020). Nanoformulations of astragalus polysaccharide with their immunomodulatory activities and antitumor effects. Shanghai, China: Shanghai University of Chinese Medicine. Master’s Thesis. 10.27320/d.cnki.gszyu.2020.000222
41
PhacharapiyangkulN.WuL. H.LeeW.-Y.KuoY. H.WuY. J.LiouH. P.et al (2019). The extracts of astragalus membranaceus enhance chemosensitivity and reduce tumor indoleamine 2, 3-dioxygenase expression. Int. J. Med. Sci.16 (8), 1107–1115. 10.7150/ijms.33106
42
QiuP.GongM.XuL. F.WangG. L.LiJ. L.XieY. (2024). Mechanism of astragalus polysaccharide in enhancing the chemosensitivity of gemcitabine in nude mice transplanted with pancreatic cancer. Chin. J. Clin. Ration. Drug Use.17 (32), 15–17. 10.15887/j.cnki.13-1389/r.2024.32.005
43
QuZ. Y. (2022). Clinical efficacy and related mechanism of astragalus polysaccharide combined with chemotherapy on advanced NSCLC. Tianjin, China: Tianjin University of Chinese Medicine. Master’s Thesis. 10.27368/d.cnki.gtzyy.2022.000129
44
ShaX.XuX.LiaoS.ChenH.RuiW. (2022). Evidence of immunogenic cancer cell death induced by honey-processed Astragalus polysaccharides in vitro and in vivo. Exp. Cell Res.410 (1), 112948. 10.1016/j.yexcr.2021.112948
45
ShenW. C.ChenS. C.WangC. H.HungC. M.PengM. T.LiuC. T.et al (2024). Astragalus polysaccharides improve adjuvant chemotherapy-induced fatigue for patients with early breast cancer. Sci. Rep.14 (1), 25690. 10.1038/s41598-024-76627-z
46
ShiW.ZhangH.TangH.FengW.ZhangZ. (2025). Effect of astragalus polysaccharide combined with cisplatin on exhaled volatile organic compounds as biomarkers for lung cancer and its anticancer mechanism. J. Pharm. Biomed. Anal.259, 116759. 10.1016/j.jpba.2025.116759
47
SongY. H.GuoZ.DouC. J.NuernisaA.ZhaoJ. (2022). Effect enhancing and toxicity-reducing activities of astragalus polysaccharide injection on U14 cervical cancer in model mice receiving X-Ray treatment. Chin. J. Exp. Tradit. Med. Formulae28 (4), 84–90. 10.13422/j.cnki.syfjx.20220421
48
SunS. Y.HeX. J.LuoD.LvC.JinS. L.LvA. P.et al (2013). Synergism and attenuation effect of Astragalus Polysaccharide on cisplatin in the treatment of mice bearing S180 tumor. J. Basic Chin. Med.19 (9), 1031–1033. 10.19945/j.cnki.issn.1006-3250.2013.09.022
49
SunL.HuangJ. X.LiX. X.ZhuoS. C.KongH. S.DuW. W.et al (2025). Astragalus polysaccharide enhances the therapeutic efficacy of cisplatin in triple-negative breast cancer through multiple mechanisms. Oncol. Res.33 (3), 641–651. 10.32604/or.2024.050057
50
TangZ.HuangG. (2022). Extraction, structure, and activity of polysaccharide from radix astragali. Biomed. Pharmacother.150, 113015. 10.1016/j.biopha.2022.113015
51
TianQ. E.LiH. D.YanM.CaiH. L.TanQ. Y.ZhangW. Y. (2012). Astragalus polysaccharides can regulate cytokine and P-glycoprotein expression in H22 tumor-bearing mice. World J. Gastroenterol.18 (47), 7079–7086. 10.3748/wjg.v18.i47.7079
52
TsaoS. M.WuT. C.ChenJ.ChangF.TsaoT. (2021). Astragalus polysaccharide injection (PG2) normalizes the neutrophil-to-lymphocyte ratio in patients with advanced lung cancer receiving immunotherapy. Integr. Cancer Ther.20, 153473542199525. 10.1177/1534735421995256
53
WagleN. S.NogueiraL.DevasiaT. P.MariottoA. B.YabroffK. R.IslamiF.et al (2025). Cancer treatment and survivorship statistics, 2025. Ca. Cancer J. Clin.75 (4), 308–340. 10.3322/caac.70011
54
WangQ. Y. (2021). Mechanism of the anti-colon cancer effects of astragalus polysaccharide based on lymphangiogenesis. Lanzhou, Gansu, China: Gansu University of Chinese Medicine. Master’s Thesis.
55
WangY. N. (2023). Regulation mechanism of astragalus polysaccharide on immune pathway of T cells in tumor microenvironment of Colon cancer. Lanzhou, Gansu, China: Gansu University of Chinese Medicine. Master’s thesis. 10.27026/d.cnki.ggszc.2023.000259
56
WangB. (2024). Professor Wang ping’s academic thoughts on treating lung cancer with the method of nurturing the primary Qi and consolidating the foundation,and study on Astragalus poIysaccharides regulating the immune checkpoints TIGIT-CD I55 to inhibit lung cancer. Tianjin, China: Tianjin University of Chinese Medicine. Master’s Thesis.
57
WangB.MaoY.YangM. (2017a). Efficacy enhancing and toxicity reducing mechanism of astragalus polysaccharides on doxorubicin therapy against mousemammary cancer. Chin. J. Hosp. Pharm.37 (6), 497–501. 10.13286/j.cnki.chinhosppharmacyj.2017.06.03
58
WangY. Q.HanY. Y.HanD. L.LinY.YangX. H. (2017b). Anti-tumor effects and mechanisms of Astragalus polysaccharide combined with cisplatin on Lewis lung cancer in mice. Shandong Med. J.57 (44), 39–41. 10.3969/j.issn.1002-266X.2017.44.011
59
WangC. H.LinC. Y.ChenJ. S.HoC. L.RauK. M.TsaiJ. T.et al (2019). Karnofsky performance status as a predictive factor for cancer-related fatigue treatment with astragalus polysaccharides (PG2) injection-a double blind, multi-center, randomized phase IV study. Cancers11 (2), 128. 10.3390/cancers11020128
60
WangK.ZhouY.LiM.ChenZ.WuZ.JiW.et al (2024). Structural elucidation and immunomodulatory activities in vitro of type I and II arabinogalactans from different origins of astragalus membranaceus. Carbohydr. Polym.333, 121974. 10.1016/j.carbpol.2024.121974
61
XuQ.ChengW.WeiJ.OuY.XiaoX.JiaY. (2023). Synergist for antitumor therapy: astragalus polysaccharides acting on immune microenvironment. Discov. Oncol.14 (1), 179. 10.1007/s12672-023-00798-w
62
YeD.ZhaoQ.DingD.MaB. L. (2023). Preclinical pharmacokinetics-related pharmacological effects of orally administered polysaccharides from traditional Chinese medicines: a review. Int. J. Biol. Macromol.252, 126484. 10.1016/j.ijbiomac.2023.126484
63
ZhangL. X. (2019). Effects of carbon ion radiation on the level of MSCs colonization related to lung cancer in tumor-bearing mice and the protective effect of astragalus polysaccharide. Lanzhou, Gansu, China: Gansu University of Chinese Medicine. Master’s Thesis. 10.27026/d.cnki.ggszc.2019.000035
64
ZhangQ.HaoS.LiL.LiuM.HuoC.BaoW.et al (2022a). M cells of mouse and human peyer’s patches mediate the lymphatic absorption of an astragalus hyperbranched heteroglycan. Carbohydr. Polym.296, 119952. 10.1016/j.carbpol.2022.119952
65
ZhangQ.LiL.HaoS.LiuM.HuoC.WuJ.et al (2022b). A lymphatic route for a hyperbranched heteroglycan from radix astragali to trigger immune responses after oral dosing. Carbohydr. Polym.292, 119653. 10.1016/j.carbpol.2022.119653
66
ZhangY.WangC.YuD.GaoY.JinH.WangY.et al (2022c). Astragalus polysaccharides improve cisplatin resistance by inhibiting EMT of lung adenocarcinoma A549/DDP cells transplanted into nude mice. Chin. J. Exp. Tradit. Med. Formulae28 (6), 79–85. 10.13422/j.cnki.syfjx.20220527
67
ZhangQ. F.SuC. Z.LuoY.ZhengF.LiangC. L.ChenY. C.et al (2024). Astragalus polysaccharide enhances antitumoral effects of chimeric antigen receptor-engineered (CAR) T cells by increasing CD122+CXCR3+PD-1- memory T cells. Biomed. Pharmacother.179, 117401. 10.1016/j.biopha.2024.117401
68
ZhouL.LiuZ.WangZ.YuS.LongT.ZhouX.et al (2017a). Astragalus polysaccharides exerts immunomodulatory effects via TLR4-mediated MyD88-dependent signaling pathway in vitro and in vivo. Sci. Rep.7 (1), 44822. 10.1038/srep44822
69
ZhouZ.MengM.NiH. (2017b). Chemosensitizing effect of astragalus polysaccharides on nasopharyngeal carcinoma cells by inducing apoptosis and modulating expression of bax/bcl-2 ratio and caspases. Med. Sci. Monit.23, 462–469. 10.12659/MSM.903170
70
ZhuZ. F.XiuQ.HanL.LiZ. H.TianY. (2017). The effect of astragalus polysaccharide on the expression of PCNA and GFAP protein in C6 glioma rat model. Int. J. Clin. Exp. Med.10, 16057–16063.
71
ZhuangM. J.LiuD.ChenY. W.MingH. X.LiY. (2017). Inhibitory effect and the mechanism of Astragalus polysaccharide combined with cisplatin on growth of inplanted Lewis lung carcinoma in mice. Chin. J. Cell Mol. Immunol.33 (04), 503–507. 10.13423/j.cnki.cjcmi.008108
Summary
Keywords
astragalus polysaccharide, preclinical studies, antitumor, meta-analysis, traditional Chinese medicine
Citation
Zhang R, Yang Q, Chen Z, Huang J and Zhang G (2025) Synergistic antitumor effects of astragalus polysaccharide: a preclinical systematic review and meta-analysis. Front. Pharmacol. 16:1672450. doi: 10.3389/fphar.2025.1672450
Received
24 July 2025
Revised
21 November 2025
Accepted
28 November 2025
Published
18 December 2025
Volume
16 - 2025
Edited by
Zhuo Luo, Guangxi Medical University, China
Reviewed by
Suryaa Manoharan, Bharathiar University, India
Saikat Sena, Lovely Professional University, India
Yu Huo, Guangxi University of Chinese Medicine, China
Updates
Copyright
© 2025 Zhang, Yang, Chen, Huang and Zhang.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Jianming Huang, wesleyhuangcn2002@163.com; Guonan Zhang, zhanggn@hotmail.com
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.