A realistic approach to evaluating the effect of baseline lipid profile in postcoronary artery bypass grafting surgery

Background: There are still many uncertainties in the association between lipid profile and postcoronary artery bypass grafting (CABG) outcomes. Although simplifying the association to linear equations makes it understandable but cannot explain many findings. Hypothesis: There is a nonlinear associatin between lipid profile indices and adverse outcomes after CABG. Methods: A total of 17 555 patients who underwent isolated CABG between 2005 and 2016 were evaluated. During the median follow-up of 75.24 months, the Restricted Cubic Splines (RCS) estimated from the Cox regression model adjusted for all possible confounders was applied to show a nonlinear relationship of lipid profile contents with the "ln hazard ratio" of mortality and major cerebro-cardiac events (MACCE). Results: The relationship between LDL-C and HDL-C with all-cause mortality was nonlinear (nonlinear p were.004 and <.001, respectively). The relationship between remnant cholesterol and all-cause mortality was linear (linearity p = .023). Among men, those in the highest LDL-C level (Q4, LDL-C > 114) and those in the lowest HDL-C level (Q1, HDL-C < 30) showed a significantly higher risk of all-cause mortality compared to other groups (compared with Q3, LDL-C Q4, HR =1.16, 95% confidence interval [CI]:1.02-1.26, p = .014; HDL-C Q1, HR =1.14, 95% CI: 1.01-1.31, p = .041). Female patients in the lowest HDL-C level (Q1, HDL-C < 30) showed a significantly higher (compared with Q3, HR = 1.14, 95% CI:1.01-1.31, p = .028) and those in the highest HDL-C level (Q4, HDL-C > 43) showed a significantly lower (compared with Q3, HR = 0.74, 95% CI:0.58-0.98, p = .019) risk of all-cause mortality. Conclusion: Determining a universal cut off for components of lipid profile may be misleading and should better be revised. Extreme values (very low or very high) for HDL-C and LDL-C have different effects on cardiovascular outcomes.