Figure 2. Gut microbiome composition is associated with markers of statin efficacy.

A) Proportion of variance explained by statin use, plasma HMG levels, and a statin-by-HMG interaction term from unadjusted PERMANOVA models or models adjusted for sex, age, BMI, and microbiome vendor using the Weighted UniFrac genus-level dissimilarity matrix. Grey area corresponds to the cumulative R-squared of variables added to the model prior to the variable indicated on the x-axis, while the colored areas of the bars represent the additional variance explained by that variable. B) Measures of gut α-diversity in statin users compared to non-users. The β-coefficient, 95%CI and p-value shown is derived from OLS models predicting each of the α-diversity measures adjusted for microbiome vendor, sex, age, and BMI. Values on the y-axis are diversity measures adjusted for these covariates (residuals). C) Measures of Observed ASVs in statin users and non-users with known therapy intensity (low, moderate, high). P-values shown correspond to β-coefficients from OLS models predicting Observed ASVs comparing each intensity group to the no statin control group, adjusted for the same covariates as in B). D) Plasma HMG levels among statin users and non-users across tertiles of gut α-diversity. Interaction P corresponds to the statin-by-α-diversity interaction term p-value from GLM predicting plasma HMG adjusted for the same covariates as in B) and C). Values on the y-axis are diversity measures adjusted for these covariates (residuals). Box plots represent the interquartile range (25th to 75th percentile, IQR), with the middle line denoting the median; whiskers span 1.5 × IQR, points beyond this range are shown individually. E) Scatter plots of Observed ASVs (x-axis) and covariate adjusted plasma HMG levels (residuals) (y-axis) in statin users with known dosage therapy intensity as well as statin non-users. Levels were adjusted for the same covariates as in B), as well as dosage intensity.