Skip to main content
. 2024 Jan 5;100:104964. doi: 10.1016/j.ebiom.2023.104964

Table 2.

Ranking of most likely causal exposures among 36 lipoproteins for BP indicators using MR-BMA.

Exposure MIP MACE (mmHg) Empirical P-value FDR
Systolic blood pressure
 S_HDL_TG 0.416 0.310 0.001 0.036
 L_LDL_TG 0.325 0.187 0.002 0.036
 L_HDL_C 0.188 −0.069 0.003 0.036
 IDL_TG 0.184 0.092 0.004 0.036
 HDL_C 0.164 −0.060 0.009 0.054
 HDL_size 0.125 −0.042 0.025 0.112
 XXL_VLDL_C 0.121 −0.058 0.009 0.054
 XS_VLDL_TG 0.117 0.057 0.011 0.057
 XL_HDL_C 0.111 −0.037 0.034 0.136
 XXL_VLDL_TG 0.105 −0.043 0.040 0.144
Diastolic blood pressure
 S_HDL_TG 0.397 0.216 0.001 0.024
 M_HDL_TG 0.302 0.105 0.002 0.024
 S_LDL_C 0.175 −0.054 0.007 0.036
 L_VLDL_C 0.161 −0.076 0.002 0.024
 LDL_C 0.145 −0.041 0.008 0.036
 XS_VLDL_TG 0.134 0.043 0.004 0.036
 Non_HDL_C 0.130 −0.036 0.007 0.036
 ApoB 0.124 −0.036 0.007 0.036
 Total_C 0.110 −0.030 0.029 0.095
 Remnant_C 0.098 −0.024 0.025 0.090

Abbreviations: BP, blood pressure; MR-BMA, Mendelian Randomisation Bayesian model averaging approach; MIP, marginal inclusion probability; MACE, model-averaged causal effect; TG, triglycerides; HDL, high-density lipoprotein; LDL, low-density lipoprotein; VLDL, very low-density lipoprotein; S_HDL_TG, small HDL TG; L_LDL_TG, large LDL TG; L_HDL_C, large HDL cholesterol; IDL_TG, intermediate-density lipoprotein TG; HDL_C, HDL cholesterol; HDL_size, average diameter for HDL particles; XXL_VLDL_C, chylomicrons and extremely large VLDL cholesterol; XS_VLDL_TG, very small VLDL TG; XL_HDL_C, very large HDL cholesterol; XXL_VLDL_TG, chylomicrons and extremely large VLDL TG; M_HDL_TG, medium HDL TG; S_LDL_C, small LDL cholesterol; L_VLDL_C, large VLDL cholesterol; LDL_C, LDL cholesterol; non_HDL_C, total cholesterol minus HDL–C; ApoB, apolipoprotein B; Total_C, Total cholesterol; Remnant_C, Remnant cholesterol (non-HDL, non-LDL-cholesterol).

Risk factors are ranked by the MIP in the primary analysis after model diagnostics. Causal effects are beta mmHg for BP indicators per 1 standard deviation increase in the risk factor. Empirical P-values are computed using 1000 permutations and adjusted for multiple testing using the false-discovery rate (FDR) procedure.