Purpose: To evaluate the effects of metformin on serum lipid profiles in nondiabetic adults through a comprehensive meta-analysis.
Methods: In the present meta-analysis, randomized and controlled trials were collected by searching PubMed, Embase, and Cochrane Libraries from inception to April 2019. Compared with placebos, the effects of metformin treatment on lipid profiles in nondiabetic adults were evaluated.
Results: Forty-seven studies from 45 articles including 5731 participants were enrolled. Pooled results showed that metformin had significant effects on total cholesterol (mean change -6.57 mg/dl; 95% CI -9.66, -3.47; P = 0.000) and LDL-c (mean change -4.69 mg/dl; 95% CI -7.38, -2.00; P = 0.001), but insignificant effects on HDL-c (mean change -4.33 mg/dl; 95% CI -9.62, 0.96; P = 0.109) and triglyceride (mean change -0.85 mg/dl; 95% CI -0.36, 2.06; P = 0.169). Significant heterogeneities were found for all lipid profiles (HDL-c = 85.5%; LDL-c = 59.9%; total cholesterol = 75.3% and triglyceride = 67.1%). Different from the pooled data, in a subgroup analysis, the effect of metformin on triglyceride in patients with polycystic ovarian syndrome (PCOS) was significant with a mean reduction of 8.15 mg/dl. In addition, sensitivity analysis showed that the pooled effects of metformin on serum lipid profiles were stable. Publication bias derived from funnel plots or Begg's tests (P = 0.933, 0.860, 0.904, and 0.567 for HDL-c, LDL-c, total cholesterol, and triglyceride, respectively) was not significant.
Conclusion: This meta-analysis revealed that metformin could reduce total cholesterol and LDL-c in nondiabetic adults. In addition, metformin might exert a triglyceride-lowering effect in nondiabetics with PCOS status.
Keywords: Meta-analysis; Metformin; Nondiabetic; Serum lipid.