Genosets for APOE and CYP7A1-rs3808607 variants do not predict LDL cholesterol lowering upon intervention with plant sterols in a randomized, double-blind, placebo-controlled trial

Abstract

Background: The consumption of 2 g/d plant sterols (PSs) reduces circulating LDL cholesterol by ≤10%. The degree of LDL cholesterol lowering was associated with specific apolipoprotein E [APOE, Reference SNP (rs)429358] and cholesterol 7α-hydroxylase (CYP7A1, rs3808607) genosets in previous post hoc analyses of randomized controlled trials. However, because post hoc analyses do not conform to the randomization model, there is a greater potential that the findings could be due to type I error, thus warranting validation through an a priori-designed intervention trial. Objectives: The GenePredict Plant Sterol study (GPS) was designed to validate associations of LDL cholesterol lowering with specific APOE and CYP7A1 genosets through a priori recruitment of individuals carrying prespecified genosets. Methods: A 2-center, double-blind, placebo-controlled, randomized 2-period crossover dietary intervention with 2 g/d PS for 28 d with a minimum 28-d washout was undertaken from July 2017 to December 2019. A priori recruitment of individuals with slightly elevated LDL cholesterol was based on genosets of APOE isoforms and CYP7A1 rs3808607. Randomization was performed with stratification by sex and genoset. Results: The recruitment target of 64 participants with prespecified genosets could not be reached, despite the screening of 477 individuals; 42 participants completed the intervention trial. Reductions in LDL cholesterol were similar across all 3 genosets (-0.298 ± 0.164, -0.357 ± 0.115, -0.293 ± 0.109 mmol/L; P = 0.0002 overall; P = 0.9126 for treatment × genoset), providing evidence that the shortfall in recruitment might not have stopped the trial from meeting the objective. Conclusions: APOE and CYP7A1 genotypes did not influence the efficacy of LDL cholesterol reductions upon dietary intervention with PSs. Findings of previous post hoc analyses could not be validated in a trial using a priori genotype-based recruitment. Obtaining adequate numbers of participants is challenging in trials using genoset-based recruitment, even for common variants.