Relationship between lipid accumulation product and new-onset diabetes in the Japanese population: a retrospective cohort study

作者：Ting Liu, Weilin Lu, Xiaofang Zhao, Tianci Yao, Bei Song, Haohui Fan, Guangyu Gao, Chengyun Liu · 发表于：Frontiers in Endocrinology · 年份：2023 · DOI：10.3389/fendo.2023.1181941 · 被引用次数：18 · 研究领域：Diabetes, Cardiovascular Risks, and Lipoproteins、Adipokines, Inflammation, and Metabolic Diseases、Cancer Research and Treatment

Background: Diabetes has become a global public health problem. Obesity has been established as a risk factor for diabetes. However, it remains unclear which of the obesity indicators (BMI, WC, WhtR, ABSI, BRI, LAP, VAI) is more appropriate for monitoring diabetes. Therefore, the objective of this investigation is to compare the strength of the association of these indicators and diabetes and reveal the relationship between LAP and diabetes. Methods: 15,252 people took part in this research. LAP was quartered and COX proportional risk model was applied to explore the relationship between LAP and new-onset diabetes. Smooth curve fitting was employed to investigate the non-linear link between LAP and diabetes mellitus. Finally, the receiver operating characteristic (ROC) curve was used to evaluate the predictive ability of the aforementioned indicators for diabetes. Results: After adjusting for confounding factors, multiple linear regression analysis showed that each unit increase in LAP was associated with a 76.8% increase in the risk of developing diabetes (HR=1.768, 95% CI: 1.139 to 2.746, P=0.011). In addition, LAP predicted new-onset diabetes better than other indicators, and the AUC was the largest [HR: 0.713, 95% CI: 0.6806-0.7454, P<0.001, in women; HR: 0.7922, 95% CI: 0.7396-0.8447; P<0.001, in men]. When LAP was used as a lone predictor, its AUC area was largest both men and women. However, after adding classical predictors (FPG, HbA1c, SBP, exercise, age) to the mode...