Extended Data Fig. 9: Variation in prevalence of the Bact2 enterotype with BMI and statin intake in the BMIS discovery cohort, and in the FGFP and CVD validation cohorts.
From: Statin therapy is associated with lower prevalence of gut microbiota dysbiosis
a–c, Variation in the prevalence of the Bact2 enterotype with BMI for statin-medicated and non-statin-medicated individuals, showing the significant effect (represented by the range bar with an asterisk; Supplementary Table 16) of statin intake given individuals’ BMI, in the BMIS obese participants (n = 474 biologically independent samples, multivariate binomial logistic regression, Statin | BMI, relative risk = 0.34, *Padj = 0.025) (a); the FGFP cohort, a population-level recruitment with a much narrower BMI range than the BMIS cohort (n = 2,345 biologically independent samples, multivariate binomial logistic regression, Statin | BMI, relative risk = 0.72, *Padj = 0.045) (b) and the MetaCardis CVD cohort (n = 271 biologically independent samples, excluding 11 individuals for which BMI was not known, multivariate binomial logistic regression, Statin | BMI, relative risk = 0.29, *Padj = 0.021) (c). In a–c, the fit lines were obtained by multinomial logistic regression of enterotypes as predicted by BMI, for statin-medicated and non-statin-medicated individuals separately, with the shaded area corresponding to the 95% confidence intervals for the Bact2 regression. Adjustment for multiple testing (Padj) was performed using the Benjamini–Hochberg method.
