Performance of Different LDL-C Equations in an Intervention Improving Atherogenic Dyslipidemia in Participants with Type 2 Diabetes

Abstract

<h3>Lead Author's Financial Disclosures</h3> S.J.A. is an employee of Virta Health and offered stock options from Virta Health. <h3>Study Funding</h3> None. <h3>Background/Synopsis</h3> LDL-C estimated using the Friedewald equation (LDL-CF) is limited in type 2 diabetes (T2D) due to hypertriglyceridemia (HTG). In our recent study of 262 patients with T2D, a low carbohydrate intervention with nutritional ketosis significantly improved atherogenic dyslipidemia (ALP) in participants with T2D, where a majority of the participants transitioned from a predominantly small, dense LDL phenotype (phenotype B) to a large, buoyant LDL phenotype (phenotype A). However, the improvement in ALP was associated with an increase in LDL-CF. <h3>Objective/Purpose</h3> This study assessed differences between LDL-CF versus other recently adopted LDL-C equations to estimate the effect of our intervention on LDL-C, and assess the performance of these equations across a broad range of HTG. <h3>Methods</h3> LDL-C were recalculated using Martin-Hopkins (LDL-CMH) and the new proposed NIH 2 (LDL-CN2) equations. The effect size of the intervention on LDL-C change from baseline to 2 years was assessed using each LDL-C equation with repeated measure ANOVA. Since LDL-CN2 is recommended for use with triglycerides (TG) up to 800mg/dL, we calculated LDL-C with this equation using two cut-offs: TG≤400mg/dL and TG≤800mg/dL. We then assessed the differences between LDL-CMH, LDL-CN2 and LDL-CF at each time point among those with TG≤400mg/dL using a linear mixed effect model (LMM) and covariates that were initially used for the assessment of LDL-CF (Table 1). <h3>Results</h3> There was an increase in LDL-C in all three LDL-C equations, with medium effect sizes in both LDL-CF (ηp2=0.08, +13.5% from 100.9mg/dL) and LDL-CN2 (ηp2 =0.07 for TG≤400mg/dL, +11.6% from 104.8mg/dL or ηp2 =0.06 for TG≤800mg/dL, +10.7% from 104.6mg/dL) and a small effect with LDL-CMH (ηp2=0.04, +9.2% from 106.5mg/dL). At baseline, LDL-CMH estimated significantly greater LDL-C than the other two equations, especially in individuals with HTG and phenotype B (Table 1). At 1 and 2 years when the proportion of individuals with ALP and HTG significantly decreased in the intervention, there were no significant differences in the estimated LDL-C between the equations. <h3>Conclusions</h3> In patients with T2D where HTG is a common lipid abnormality, LDL-CMH may be the most appropriate equation to use in assessing change in LDL-C. For patients with greater than 400mg/dL of TG, LDL-CN2 can be alternatively considered. Future studies should compare these equations with direct LDL-C measurements in patients with T2D.