Association of atherogenic index of plasma with cardiovascular disease mortality and all-cause mortality in the general US adult population: results from NHANES 2005–2018

Abstract Background The atherogenic index of plasma (AIP) is a critical metric for predicting cardiovascular outcomes. However, its associations with cardiovascular disease mortality (CVM) and all-cause mortality (ACM) remain unclear. This study aims to elucidate the relationship between baseline AIP levels and CVM and ACM among a broad cohort of US adults. Methods Utilizing data from the National Health and Nutrition Examination Survey (2005–2018), we analyzed 18,133 adults aged ≥ 18. Baseline triglycerides and high-density lipoprotein cholesterol levels were measured to calculate the AIP. Mortality outcomes were determined through linkage with the National Death Index database, with follow-up through December 31, 2019. Multivariable Cox proportional hazard models examined the associations between baseline AIP and mortality risks. Additionally, restricted cubic splines were utilized to investigate potential non-linear relationships, with subgroup analyses conducted across strata defined by age, gender, body mass index, diabetes, hypertension, and metabolic syndrome to assess variability in these associations. Results Over a median 95.0-month follow-up, there were 1870 all-cause deaths and 579 cardiovascular disease-related deaths. Our findings indicate a J-shaped association between the AIP and ACM (threshold = 0.0905); specifically, when baseline AIP exceeded 0.0905, a significant positive association with ACM emerged (hazard ratio, HR (95% confidence interval, CI): 1.61(1.08–2.37)). However, after adjusting for confounders, the relationship between AIP and CVM was not statistically significant (HR 1.31, 95% CI 0.93–1.86). Notably, in the 40–60-year age group, AIP was significantly positively associated with ACM and CVM, with HRs and 95% CIs of 1.51 (1.08v2.10) and 2.63 (1.39–4.98), respectively. Conclusions A J-shaped relationship was observed between baseline AIP levels and ACM within the general US population, with a threshold of 0.0905. Moreover, AIP could potentially be an effective predictor for future ACM or CVM, particularly among individuals aged 40–60. Further investigation is warranted to corroborate these findings.