系统评价黄连素成分单用或联合口服降糖药辅助治疗T2DM的降糖疗效与安全性，并探讨不同基线血糖水平对效应的一致性影响。

系统检索PubMed、Web of Science、Cochrane Library、知网、万方和维普数据库等数据库建库至2026年1月1日，纳入T2DM临床研究，干预为黄连素单用或联合其他常规降糖干预措施，提取FPG、HbA1c、2hPBG及总不良事件与低血糖结局，并进行基线FPG或HbA1c分层亚组分析，采用漏斗图与Egger检验评估发表偏倚。

共纳入33项研究、3131例患者，干预持续14天至6个月，黄连素常用剂量0.9~2.4 g/d。与对照相比，黄连素可显著降低FPG（WMD=-0.78 mmol/L）、HbA1c（WMD=-0.63%）及2hPBG（MD=-1.11 mmol/L）。按基线分层后，各亚组效应方向总体一致，其中HbA1c≥9.0%人群降幅更明显（-0.79% vs -0.58%）。安全性方面，研究组总不良事件风险降低（RR=0.73），低血糖风险亦降低（RR=0.59）。漏斗图与Egger检验未提示主要结局存在显著小样本效应。

虽然黄连素并不能作为单一有效成分直接治疗T2DM，但作为T2DM的辅助干预可带来额外降糖获益，并未增加安全性负担，且可能降低总不良事件与低血糖风险，基线血糖控制水平可能影响效应一致性。未来仍需进一步开展更高质量随机对照试验，以进一步界定最获益人群与最佳应用场景。

To systematically evaluate the efficacy and safety of berberine alone or combined with oral hypoglycemic agents in type 2 diabetes mellitus (T2DM), and explore whether baseline glycemic status affects treatment effect consistency.

We systematically searched PubMed, Web of Science, Cochrane Library, CNKI, Wanfang, and VIP databases from inception to January 1, 2026. We included T2DM clinical studies using berberine alone or combined with standard glucose-lowering therapies. Outcomes included fasting plasma glucose (FPG), glycated hemoglobin (HbA1c), 2-hour postprandial blood glucose (2hPBG), and safety (adverse events, hypoglycemia). Subgroup analyses were performed by baseline FPG/HbA1c; publication bias was assessed via funnel plots and Egger’s test.

Thirty-three studies (3,131 participants) were included, with intervention durations of 14 days to 6 months and berberine doses of 0.9-2.4 g/day. Compared with controls, berberine significantly reduced FPG (WMD=-0.78 mmol/L), HbA1c (WMD=-0.63%), and 2hPBG (MD=-1.11 mmol/L). Subgroup effects were directionally consistent; HbA1c reduction was greater in patients with baseline HbA1c≥9.0% (-0.79% vs -0.58%). Berberine was associated with lower risks of total adverse events (RR=0.73) and hypoglycemia (RR=0.59). No significant publication bias was detected.

Berberine as adjunctive therapy improves glycemic control in T2DM without increasing safety risk, and may reduce adverse events and hypoglycemia. Baseline glycemic status may affect effect magnitude. Further high-quality randomized controlled trials are needed to confirm optimal populations and clinical applications.