This paper introduces the concept of multivariable Mendelian randomization, an extension of traditional Mendelian randomization that uses multiple genetic variants associated with multiple risk factors to estimate the causal effects of each risk factor on an outcome. The authors propose this approach to address scenarios where finding a single genetic variant solely associated with a specific risk factor is challenging, such as in the case of triglyceride levels and cardiovascular disease. They present methods for estimating causal effects using both individual-level and summarized data, including 2-stage least squares (2SLS) and likelihood-based approaches. The paper also discusses the assumptions necessary for a valid multivariable Mendelian randomization analysis and demonstrates the application of these methods through an example and a simulation study. The example focuses on the causal effects of low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and triglycerides on coronary heart disease (CHD) risk. The simulation study evaluates the statistical properties of the methods under different scenarios, including the presence of causal relationships between risk factors. The authors conclude that multivariable Mendelian randomization can provide valuable insights into the causal effects of risk factors, even when no variant is uniquely associated with a specific risk factor, but they also highlight the limitations and assumptions that must be considered.This paper introduces the concept of multivariable Mendelian randomization, an extension of traditional Mendelian randomization that uses multiple genetic variants associated with multiple risk factors to estimate the causal effects of each risk factor on an outcome. The authors propose this approach to address scenarios where finding a single genetic variant solely associated with a specific risk factor is challenging, such as in the case of triglyceride levels and cardiovascular disease. They present methods for estimating causal effects using both individual-level and summarized data, including 2-stage least squares (2SLS) and likelihood-based approaches. The paper also discusses the assumptions necessary for a valid multivariable Mendelian randomization analysis and demonstrates the application of these methods through an example and a simulation study. The example focuses on the causal effects of low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and triglycerides on coronary heart disease (CHD) risk. The simulation study evaluates the statistical properties of the methods under different scenarios, including the presence of causal relationships between risk factors. The authors conclude that multivariable Mendelian randomization can provide valuable insights into the causal effects of risk factors, even when no variant is uniquely associated with a specific risk factor, but they also highlight the limitations and assumptions that must be considered.