Table 2. Sex is a Significant Effect Modifier.

Sex-by-Statin Interaction Term Significance.

	All (Unrestricted by Baseline Aggression)
Male x statin interaction term
	Statin vs placebo: N = 970	Simva vs placebo: N = 646	Prava vs placebo: N = 647
Beta (SE)	-1.2 (0.45)	-1.4 (0.54)	-0.93 (0.53)
95%CI	-2.1, -0.33	-2.5, -0.38	-2.0, 0.11
P-value	0.007	0.008	0.079
	No Baseline Aggression (Baseline OASMa = 0)
Male x statin interaction term
	Statin vs placebo: N = 514	Simva vs placebo: N = 344	Prava vs placebo: N = 364
Beta (SE)	-1.3 (0.38)	-1.1 (0.46)	-1.4 (0.50)
95%CI	-2.0, -0.53	-2.0, -0.21	-2.4, -0.45
P-value	0.001	0.015	0.004

Beta = regression coefficient; CI = confidence interval; OASMa = Overt-Aggression-Scale-Modified–Aggression-Subscale; Prava = pravastatin; SE = standard error; Simva = simvastatin. Regressions (robust SEs) adjust for baseline OASMa, as a primary source of change variance; the sex-interaction, and (as required) the individual components of the interaction term. Excludes outliers (OASMa absolute change >40 points). Without exclusion of outliers: Results lose significance for the full group. Results remain identical for the No Baseline Aggression group (all aggression outliers had nonzero baseline aggression).