Efficacy and safety of low levels of low-density lipoprotein cholesterol: trans-ancestry linear and non-linear Mendelian randomization analyses

Lay Summary We used the Mendelian randomization method to estimate the causal relationships between LDL-C and efficacy and safety outcomes in the UK Biobank and the China-PAR project.We found a linear rather than a non-linear relationship between genetically proxied LDL cholesterol and coronary artery disease.There are potential safety concerns (including haemorrhagic stroke and dementia) for people who have low LDL-C levels. Aims LDL cholesterol (LDL-C) is a well-established risk factor for coronary artery disease (CAD). However, the optimal LDL-C level with regard to efficacy and safety remains unclear. We aimed to investigate the causal relationships between LDL-C and efficacy and safety outcomes. Methods and results We analyzed 353 232 British from the UK Biobank and 41 271 Chinese from the China-PAR project. Linear and non-linear Mendelian randomization (MR) analyses were performed to evaluate the causal relation between genetically proxied LDL-C and CAD, all-cause mortality, and safety outcomes (including haemorrhagic stroke, diabetes mellitus, overall cancer, non-cardiovascular death, and dementia). No significant non-linear associations were observed for CAD, all-cause mortality, and safety outcomes (Cochran Q P > 0.25 in British and Chinese) with LDL-C levels above the minimum values of 50 and 20 mg/dL in British and Chinese, respectively. Linear MR analyses demonstrated a positive association of LDL-C with CAD [British: odds ratio (OR) per unit mmol/L increase, 1.75, P = 7.57 x 10(-52); Chinese: OR, 2.06, P = 9.10 x 10(-3)]. Furthermore, stratified analyses restricted to individuals with LDL-C levels less than the guideline-recommended 70 mg/dL demonstrated lower LDL-C levels were associated with a higher risk of adverse events, including haemorrhagic stroke (British: OR, 0.72, P = 0.03) and dementia (British: OR, 0.75, P = 0.03). Conclusion In British and Chinese populations, we confirmed a linear dose-response relationship of LDL-C with CAD and found potential safety concerns at low LDL-C levels, providing recommendations for monitoring adverse events in people with low LDL-C in the prevention of cardiovascular disease.