Background <p>The association of the triglyceride–glucose index, a validated marker of insulin resistance, with specific dietary patterns is unclear.</p> Methods <p>This cross-sectional analysis of NHANES 1999–2018 included 19,016 adults. Logistic regression, restricted cubic splines, and interaction analyses were used to evaluate associations between macronutrient intake and the triglyceride–glucose index. This analysis was designed a priori to evaluate whether a low-carbohydrate, high-fat macronutrient distribution (captured by PEC, PEF, and carbohydrate-to-fat energy ratio) is associated with the TyG index.</p> Results <p>Higher quartiles of percentage energy from carbohydrates and carbohydrate-to-fat energy ratio were associated with higher odds of an elevated TyG index, whereas higher quartiles of percentage energy from fat were associated with lower odds. Restricted cubic splines showed a positive linear association for carbohydrate energy and an inverse linear association for fat energy with TyG. Interaction analyses supported that a low-carbohydrate, high-fat pattern correlated with lower fasting triglyceride levels, and mediation analysis indicated that fasting triglycerides explained ~ 49% of the association between carbohydrate-to-fat energy ratio and TyG.</p> Conclusions <p>A low-carbohydrate, high-fat diet was significantly associated with a lower triglyceride–glucose index; this association was primarily mediated by reductions in fasting triglyceride levels.</p> Graphical Abstract <p></p>