Comparing the pharmacological effects of different ginsenosides in rodent models of heart failure: a systematic review and network meta-analysis

Objective To evaluate and compare the effects and potential advantages of different ginsenosides in animal models of heart failure (HF). Methods We systematically searched for and extracted data from studies of ginsenosides, including Rb1, Rg1, Rg3 and Rg5, and total ginsenosides in animal models of HF. Outcomes included left ventricular ejection fraction (LVEF), left ventricular fractional shortening (LVFS), left ventricular end-diastolic pressure (LVEDP), maximum rate of left ventricular pressure rise (+dP/dtmax), serum brain natriuretic peptide (BNP) and the heart weight index. Random-effects network meta-analysis estimated effects relative to model controls, with 95% confidence intervals (CIs) and 95% prediction intervals (PrIs). All networks were star-shaped, sharing the model control as a common comparator, and lacked closed loops; statistical inconsistency testing was therefore not possible. Exploratory analyses examined effect modification by model type, animal species, administration route and treatment duration. Sensitivity analyses included stratification by model type, separate analyses of BNP and NT-proBNP, alternative effect scales for HW/BW, and networks restricted to individual ginsenosides. Study quality was assessed using the SYRCLE risk-of-bias tool. The study sought to compare the pharmacological profiles of individual ginsenosides and provide evidence to inform intervention selection for clinical trials in HF. Results We included 29 studies covering 10 ginsenoside interventions. Several ginsenosides significantly improved cardiac functional outcomes compared with HF model controls. For serum BNP reduction, Rg1 (SMD = −2.38, 95% CI -4.51 to −0.25), Rb1 (SMD = −2.60, 95% CI -3.80 to −1.40) and total ginsenosides (SMD = −1.94, 95% CI -3.07 to −0.82) showed significant effects. However, the Rg1 effect was no longer significant after mixture nodes were excluded. Significant increases in LVEF were observed with total ginsenosides (MD = 17.56, 95% CI 7.65–27.48), Re (MD = 17.31, 95% CI 2.16–32.46), Rg3 (MD = 16.09, 95% CI 6.03–26.14), Rb1 (MD = 12.09, 95% CI 6.71–17.46) and Rg1 (MD = 13.68, 95% CI 3.28–24.09). LVFS was significantly higher with Rg1 (MD = 11.72, 95% CI 4.92–18.52), Rg3 (MD = 9.90, 95% CI 5.50–14.29), total secondary ginsenosides (MD = 10.30, 95% CI 2.12–18.48), Rb1 (MD = 8.95, 95% CI 6.16–11.74),Re (MD = 8.29, 95% CI 0.24–16.34) and Rd (MD = 6.48, 95% CI 0.45–12.51). Total ginsenosides (MD = −3.97, 95% CI -4.39 to −3.55), Re (MD = −1.65, 95% CI -2.24 to −1.06), Rb1 (MD = −1.49, 95% CI -2.54 to −0.44) and Rb2 (MD = −1.40, 95% CI -2.48 to −0.32) significantly reduced LVEDP. Total ginsenosides (MD = 1374.00, 95% CI 425.00–2,323.00) and Rb1 (MD = 1200.74, 95% CI 718.89–1682.60) significantly increased +dP/dtmax. For the heart weight-to-body weight (HW/BW) ratio, none of the five nodes (Rh2, Rb1, Rg5, Re and Rg1) showed a significant effect (all P > 0.05). The available evidence was insufficient to determine the effects of ginsenosides on HW/BW. Conclusion Some ginsenosides improve haemodynamic and cardiac functional outcomes in rodent models of HF. Given the methodological limitations of the included studies, these findings require confirmation in high-quality preclinical studies. Systematic review registration https://osf.io/ye48s/ ; Identifier, YE48S