Exploring the association between vitamin D levels and dyslipidemia risk: insights from machine learning and generalized additive models.

INTRODUCTION

Vitamin D is a necessary nutrient that is important for calcium homeostasis and bone health. Dyslipidemia is thought to be a risk factor for the development of atherosclerotic illnesses. Recent research suggests that vitamin D may influence lipid metabolism, specifically the levels of circulating lipids in the blood. However, the relationship between vitamin D and dyslipidemia remains controversial, indicating a need for further research to clarify this association.

OBJECTIVES

Data from 780 participants in the "Early Identification, Early Diagnosis Techniques, and Points of Risk for Diabetes in Major Chronic Non-communicable Disease Prevention and Control Studies" were analyzed.

METHODS

We employed machine learning with the XGboost algorithm, Least Absolute Shrinkage Selection Operator (LASSO) regression, and univariate logistic regression to screen variables. Subsequently, multiple logistic regression and a generalized additive model (GAM) were utilized to construct models analyzing the association between vitamin D levels and dyslipidemia.

RESULTS

In our study, the XGboost machine learning algorithm explored the relative importance of all included variables, confirming a robust association between vitamin D levels and dyslipidemia. After adjusting for all the important covariates, the results showed that the risk of dyslipidemia in vitamin D insufficiency group and vitamin D deficiency group was 2.11 times and 2.77 times of that in vitamin D sufficiency group, respectively. A smooth curve was constructed based on GAM and a significant negative association was found between 25(OH)D and the risk of dyslipidemia.

CONCLUSION

There may be a negative association between 25(OH)D and the risk of dyslipidemia. Nonetheless, additional well-designed studies are necessary to substantiate this relationship.