Causal relationship between consumption of different tea types and the risk of stroke: A two-sample Mendelian randomization analysis

This study aims to investigate whether the genetically predicted tea consumption of different types is causally associated with the risk of stroke. Four statistical datasets from genome-wide association studies (GWASs) related to tea consumption were employed. Seven distinct stroke types sourced from the European Bioinformatics Institute database were designated as outcomes. Two-sample univariate and multivariate Mendelian randomization (MVMR) analyses respectively were conducted to assess the causal relationship between tea consumption of different types and stroke risk. Two-sample MR analyses were conducted using 1061 single nucleotide polymorphisms as instrumental variables across seven stroke types. Univariate analyses showed that Rooibos tea intake was associated with large artery atherosclerosis stroke (LAAS; OR = 1.096, 95% CI= 1.030-1.165, P = .004), small-vessel stroke (OR = 1.084, 95% CI= 1.021-1.151, P = .008), lacunar stroke (OR = 1.055, 95% CI= 1.002-1.111, P = .042), and cardioembolic stroke (CES; OR = 1.059, 95% CI= 1.012-1.107, P = .013). The directional discrepancies between MR-Egger and inverse variance weighting analyses (suggesting horizontal pleiotropy) cast doubt on the MR-derived causal links between Rooibos tea intake and LAAS/CES, implying potential pleiotropic bias over true causality. Further multivariate analyses supported the causal relationship of rooibos tea intake with a reduced risk of LAAS (OR = 0.942, 95% CI= 0.888-0.999, P = .049) and CES (OR = 0.952, 95% CI= 0.910-0.996, P = .033). No significant horizontal pleiotropy was detected, indicating the robustness of the results. This study substantiated the causal relationship between tea consumption and stroke risk. Rooibos tea intake may be causally associated with a reduced risk of LAAS and CES.