Association between triglyceride glucose-body mass index and outcomes in patients with acute ischemic stroke: a retrospective secondary analysis of a prospective Korean cohort

Background The triglyceride glucose-body mass index (TyG-BMI) has recently garnered attention as a robust surrogate indicator for insulin resistance (IR). However, despite this burgeoning interest, the prospective effect of TyG-BMI on adverse clinical outcomes in acute ischemic stroke (AIS) individuals remains incompletely understood. Consequently, this investigation seeks to elucidate the effect of TyG-BMI on 90-day poor outcomes in AIS individuals. Methods During the period of January 2010 to December 2016, this longitudinal investigation encompassed 1,722 individuals diagnosed with AIS who were treated at Seoul National University Hospital (monocentric) in Korea. Blood samples were collected at admission, and all patients were admitted within 7 days after stroke onset. Binary logistic regression analysis was subsequently utilized to evaluate the effect of TyG-BMI on adverse outcomes observed at 90 days. Potential nonlinear dose–response associations were employed using restricted cubic spline (RCS) models to identify threshold (inflection) effects. Results The final analysis included a total of 1,722 individuals, with males constituting 61.56% of the cohort. Among these participants, 393 (22.82%) were below the age of 60, 457 (26.54%) were aged between 60 and 70 years, 610 (35.42%) fell within the 70 to 80-year age bracket, and 262 (15. 21%) were over 80 years of age. The median admission NIHSS score was 3 (1-7). In multivariable logistic regression analyses, TyG-BMI failed to exhibit any meaningful linear correlation with 90-day poor outcome in the overall AIS cohort. RCS analyses, however, identified a nonlinear effect of TyG-BMI on 90-day poor outcome. An inflection point in this association was determined at a TyG-BMI value of 193.59. For TyG-BMI values less than or equal to 193.59, a negative effect of TyG-BMI on 90-day poor outcomes (OR = 0.65, 95% CI: 0.48–0.89). In contrast, for TyG-BMI values exceeding this threshold, the effect of TyG-BMI on 90-day unfavorable outcomes was not statistically significant (OR = 1.16, 95% CI: 0.93–1.45, p = 0.1765). Conclusion Our findings indicate a nonlinear relationship and threshold effect between the TyG-BMI and 90-day unfavorable outcomes in patients with AIS. Only when TyG-BMI ≤ 193.59, the TyG-BMI suggested a negative association with 90-day unfavorable outcomes.