Gene–diet interaction between CETP rs3764261 and vegetarian diet on HDL-C levels: evidence for precision nutrition from the Taiwan Biobank
Background Precision nutrition seeks genotype-informed dietary guidance. We examined the associations of CETP rs3764261 and a vegetarian diet with high-density lipoprotein cholesterol (HDL-C) levels and the potential modification of this association by the variant in Taiwanese adults. Methods We analyzed 9,263 adults from the Taiwan Biobank with rs3764261 genotype data and available HDL-C measurements. Rs3764261 genotypes were obtained from whole-genome genotyping using the TWBv2.0 Affymetrix SNP array, with quality control based on Hardy–Weinberg equilibrium, minor allele frequency, and call-rate thresholds. Vegetarian diet status was defined using questionnaire data (being vegetarian throughout the day for >6 months). Associations were evaluated using multiple linear regression models, adjusting for relevant covariates and testing the rs3764261 × vegetarian-diet interaction. Results A total of 9,263 adults were included, comprising 8,330 non-vegetarians and 933 vegetarians. Vegetarians had significantly lower mean HDL-C than non-vegetarians (53.190 vs. 56.031 mg/dL, p < 0.0001) and lower LDL-C (p < 0.0001). The distribution of CETP rs3764261 genotypes did not differ by diet (p = 0.8413). In multivariable analysis, rs3764261 CA (β = 3.3318, p < 0.0001) and AA (β = 8.4651, p < 0.0001) genotypes were associated with higher HDL-C compared with CC. However, a vegetarian diet was associated with lower HDL-C (β = −3.4040, p < 0.0001). A significant interaction between rs3764261 and vegetarian diet was observed (p for interaction = 0.0045). The positive association of CA and AA genotypes with HDL-C was evident among non-vegetarians, but not vegetarians, and the HDL-lowering effect of a vegetarian diet became progressively stronger across the CC, CA, and AA genotypes (β = −2.6779, −4.8479, and −8.1884 mg/dL, respectively). However, the small number of AA homozygotes among vegetarians (n = 21) limited statistical power for some stratified estimates, and these findings warrant replication in independent cohorts. Conclusion Rs3764261 modified the association between vegetarian diet status and HDL-C levels, supporting evidence for gene–diet interactions in precision nutrition approaches to lipid risk management.