Non-linear relationship between the atherogenic index of plasma and ischemic stroke in the diabetic population in ICU: A multicenter retrospective cohort study

Objectives The Atherogenic Index of Plasma (AIP) value is relationship with the risk of atherosclerosis, a known risk factor for cardiovascular events. However, studies on the correlation between AIP value and ischemic stroke (IS) in the diabetic population in Intensive Care Unit (ICU) are rare. Our study aimed to investigate the relationship between AIP values and IS among diabetic patients in American ICUs. Methods A multicenter retrospective cohort study comprising 3695 patients from the eICU-CRD v2.0 database between 2014 and 2015 in the USA was conducted. We utilized logistic regression model to investigate the correlation between between AIP values and IS among diabetic patients in American ICUs. To detect possible non-linear associations, we combined logistic regression with generalized additive model (GAM). Additionally, we conducted a thorough array of sensitivity and subgroup analyses to verify the robustness of our results. Results The prevalence of IS was 19.46%. The median AIP was 0.52 (interquartile range, 0.29-0.76). Participants with stroke exhibited a significant elevation in AIP levels. In particular, each one-unit elevation in AIP levels was associated with a 40% increased risk of IS (OR=1.4, 95% CI 1.1-2.8, P<0.001). In addition, a non-linear relationship exists between the AIP value and the incidence of IS, with an inflection point at 0.8. The effect sizes (OR) on the left and right sides of the inflection point were 0.9 (95%CI: 0.5-1.8; P= 0.741) and 1.7 (95%CI: 1.2-2.5; P＜0.001 ), respectively. Conclusion The research uncovers a positive, non-linear correlation between the AIP value and IS incidence among diabetic patients in American ICUs. Notably, a significant association between the AIP value and IS emerges when the AIP value is less than 0.8. From a therapeutic perspective, reducing AIP levels below the inflection point seems reasonable. However, the findings require validation through prospective studies.