Statin safety in prevention of cardiovascular diseases: causal inference and risk prediction

Background: The widespread concerns about statin safety have resulted in low uptake of and poor adherence to statin treatment for prevention of cardiovascular diseases. The use of statins for primary prevention has been particularly challenging due to the controversy about the balance between benefits and harms of treatment. Personalised clinical decision-making and stratified treatment strategies that take into account the risk of adverse events are potential approaches towards better use of statins. Methods: A systematic review of randomised controlled trials was conducted, with pair-wise, network, and dose-response meta-analyses, to assess the associations between statins and common adverse events and explore the variations by drug type and dose in primary prevention patients. A prognostic model (StatinMD) was derived and externally validated to predict the personalised risk of serious muscle disorders in individuals eligible for statin treatment, using a competing risk model with data from electronic healthcare records. Results: Statins were associated with a small increase in the risk of muscle symptoms, liver dysfunction, renal insufficiency, and eye conditions, but not with muscle disorders or diabetes. There was little evidence of the difference between statin drugs or the dose-response relationships of their adverse effects. The StatinMD model included 22 predictors to predict the risk of serious muscle disorders in 1, 5, and 10 years. The model showed overall good discrimination and calibration in the majority of the population. Conclusions: The overall balance between benefits and harms of statins supports their use for primary prevention of cardiovascular diseases. The StatinMD model provides a reliable predicted risk of serious muscle disorders for most individuals to assist clinical decision-making on statin treatment.