AUTHOR=Zhang Yanan , Wang Ruiqing , Tang Xinhua , Wang Yanjun , Guo Ping , Wang Shukang , Liu Jing
TITLE=A Mendelian Randomization Study of the Effect of Tea Intake on Type 2 Diabetes
JOURNAL=Frontiers in Genetics
VOLUME=13
YEAR=2022
URL=https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2022.835917
DOI=10.3389/fgene.2022.835917
ISSN=1664-8021
ABSTRACT=Background: The association reported between tea intake and type 2 diabetes (T2D) is inconsistent in previous studies and remains controversial. We aimed to explore the causal relationship between tea intake, T2D, and glycemic traits including hemoglobin A1c (HbA1c), fasting plasma glucose (FPG), fasting serum insulin (FSI), and homeostasis model of insulin resistance (HOMA-IR) levels.
Methods: A 2-sample Mendelian randomization (MR) was performed using summary statistics from large-scale genome-wide association studies of tea intake from the UK Biobank, T2D from the DIAGRAM consortium, and glycemic traits from the Magic consortium. The findings were verified through sensitivity analyses using various MR methods with different model assumptions and by comprehensively evaluating the influence of pleiotropy effects and outliers.
Results: With the use of a two-sample MR with inverse variance-weighted method, the odds ratio per unit SD change of tea intake (SD: 2.85 cups/day) for T2D, HbA1c, FPG, FSI, and HOMA-IR levels was 0.949 (95% CI 0.844–1.067, p = 0.383), 0.994 (95% CI 0.975–1.013, p = 0.554), 0.996 (95% CI 0.978–1.015, p = 0.703), 0.968 (95% CI 0.948–0.986, p = 0.001), and 0.953 (95% CI 0.900–1.009, p = 0.102), respectively. The results were consistent with those of the other six methods that we used with different model assumptions, suggesting that the findings were robust and convincing. We also performed various sensitivity analyses for outlier removal, pleiotropy detection, and leave-one-out analysis.
Conclusion: Our MR results did not support the causal effect of tea intake on T2D and crucial glycemic traits. These findings suggest that previous observational studies may have been confounded.