부산대학교 도서관

Introduction:Wide variability in LDLC change is observed with statins, yet determinants of statin response are uncertain.Hypotheses:Some markers of inflammation are associated with decreased statin response. Components of metabolic syndrome and older age are associated with better response.Methods:JUPITER (NCT00239681) randomized participants with LDLC<130 mg/dL and hs-CRP≥2 mg/L to rosuvastatin 20 mg/d or placebo. Baseline and 1-year levels of LDLC and a panel of 20 other biomarkers were measured. We defined optimal statin response as ≥50% LDLC reduction; suboptimal response as <50% reduction; and non-response as no change or increase in LDLC. χ2, t-tests and ANOVA were used to compare variables across optimal (N=3,801), suboptimal (N=3,235) and non-responders (N=868). Multinomial regression models evaluated associations of determinants of statin response after adjusting for risk factors including compliance. Backward selection identified variables that associated with response. t-distributed Stochastic Neighbor Embedding (t-SNE) suggested predictor clusters for statin response.Results:Compliance was the strongest determinant of statin response (Table). On multivariable analyses, blacks and Hispanics were at higher risk for suboptimal/non-response compared with whites as were greater alcohol use and HbA1c. Higher baseline levels of lipids [lipoprotein(a), apolipoprotein (apo) AI and B], inflammatory biomarkers, and lower branched-chain amino acids associated with decreased statin response (Table). By contrast, determinants of better response included older age, female sex, and higher waist circumference and baseline levels of LDLC, triglycerides, and blood pressure. We observed systematic patterns in t-SNE clustering for statin response by LDLC, LDL particle number, and apo B.Conclusions:This study identified determinants for high-intensity statin response and suggests other pathways of CVD risk beyond those addressed by statin treatment that require further investigation.