**Background:** Diabetes affects 10% of the global population and is associated with complications such as retinopathy, nephropathy, and cardiovascular disease. Natural products with multi-target hypoglycemic effects are promising alternatives to synthetic drugs. *Engelhardia roxburghiana* Wall leaves (LERW) have been used as sweet tea in Chinese folk medicine and are rich in flavonoids, particularly astilbin. While astilbin has known hypoglycemic activity, the comparative efficacy of the whole extract versus the purified compound had not been established. This study aimed to compare the antioxidant and hypoglycemic activities of ethanol extract of LERW (E-LERW) and astilbin using in vitro assays and an in vivo diabetic mouse model.
**Methods:** E-LERW was prepared by microwave-assisted extraction with 60% ethanol. Composition was analyzed by HPLC-MS/MS, and total flavonoids, total phenols, and astilbin content were quantified. Antioxidant activity was assessed via DPPH and ABTS radical scavenging, ferric reducing power (FRAP), and inhibition of lipid membrane oxidation. α-Glucosidase inhibition was measured using pNPG substrate, and kinetic analysis (Lineweaver-Burk plots) determined inhibition type and Ki values. For in vivo study, male Kunming mice were injected with alloxan (200 mg/kg) to induce diabetes. Diabetic mice (blood glucose >11.1 mmol/L) were divided into 6 groups (n=6): model control (MC), astilbin control (AC, 30 mg/kg), metformin positive control (PC, 100 mg/kg), and E-LERW high (H, 600 mg/kg), medium (M, 300 mg/kg), and low (L, 150 mg/kg) doses. Treatments were administered orally twice daily for 28 days. Fasting blood glucose (FBG), oral glucose tolerance (OGTT), insulin, triglycerides (TG), total cholesterol (TC), HDL, LDL, body weight, food/water intake, excretion, and liver/kidney organ indexes were measured.
**Key Results:** HPLC-MS/MS identified 10 components in E-LERW, with astilbin as the predominant compound. E-LERW contained 94.79 ± 2.49 mg/g astilbin, 153.42 ± 2.74 mg RE/g total flavonoids, and 255.74 ± 4.16 mg GE/g total phenols. In antioxidant assays, E-LERW showed significantly higher activity than astilbin (p < 0.05) in all tests; at 2 mg/mL, E-LERW inhibited 75% of lipid membrane oxidation versus <20% for astilbin. For α-glucosidase inhibition, IC50 values were 0.46 ± 0.09 mg/mL for E-LERW, 1.12 ± 0.17 mg/mL for astilbin, and 0.19 ± 0.03 mg/mL for acarbose. Kinetic analysis revealed mixed-type inhibition with Ki values of 0.145 mg/mL (E-LERW) and 0.474 mg/mL (astilbin). In vivo, alloxan injection increased FBG to 4.18 times normal, and insulin dropped to 4.72% of normal. After 28 days, E-LERW (M) reduced FBG by 16.64% compared to MC, and increased insulin by 494.52% (to 28.0% of normal). E-LERW (M) also reduced TG by 12.87%, TC by 32.70%, and LDL by 22.99%, while increasing HDL. Food intake, water intake, and excretion decreased by 27.29%, 36.15%, and 30.93%, respectively, with E-LERW (M). Liver index increased 33% in MC but only 21% with E-LERW (M); kidney index increased 67% in MC versus 44% with E-LERW (M). Compared to astilbin control, E-LERW (M) showed similar glucose-lowering but superior effects on insulin secretion, organ protection, and reduction of water intake and excretion.
**Clinical Implications:** E-LERW demonstrates potent antioxidant and hypoglycemic activities, likely due to synergistic effects of astilbin and other polyphenols. The extract's ability to enhance insulin secretion and protect pancreatic islets and organs suggests it may be a promising functional ingredient for diabetes management. While astilbin appears to be the primary glucose-lowering component, the whole extract offers additional benefits in preserving β-cell function and reducing diabetic complications. These findings support the traditional use of LERW as sweet tea and its potential development as a nutraceutical for diabetes adjuvant therapy. Further clinical studies are warranted to confirm efficacy and safety in humans.