Predictors of statin intolerance in primary care using real-world data

ObjectiveThe objective of this study was to explore the predictors of statin intolerance in the primary and secondary prevention of CVD among patients in the first two years after the date of first prescription using real-world data. MethodsThis study used the Electronic Practice Based Research Network Linked Dataset. An algorithm, which considered the muscle symptoms and creatinine kinase of patients, was used to identify statin intolerant patients. The R software was used for all analyses. Descriptive and multivariate logistic regression analyses were performed along with sensitivity analysis which was done using the Akaike Information Criterion model selection method. ResultsOverall, 4,016 patients accounting for 60,873 visits met the selection criteria. About 3.5% of the patients were statin intolerant. After adjusting for all other variables, statin intolerance was positively associated with gender (AOR 1.5, 95% CI 1.0 - 2.2), SEIFA index (AOR 3.8, 95% CI 2.3 - 6.7), employment status (AOR 2.4, 95% CI 1.1 - 5.7), and comorbidities (AOR 7.0, 95% CI 2.2 - 19.0). A similar direction of associations was seen for the exposures of the model from the sensitivity analysis and the regression model. However, since the unrecorded employment status showed a positive association, the sensitivity analysis suggests that the relationship may be influenced by residual confounding or information bias, indicating that this finding should be interpreted with caution. ConclusionStatin intolerance within the diverse community represented in the dataset is driven by gender, employment status, area-based social advantage and disadvantage index, and comorbidities.