The aim of the Probucol Quantitative Regression Swedish Trial (PQRST) (n = 303) was to investigate whether probucol (0.5 g BID) added to diet and cholestyramine (8 g BID) could retard progression or induce regression of femoral atherosclerosis in hypercholesterolemic (> 6.86 mmol/L) subjects. Probucol did not induce regression over the 3-year trial period as estimated by change in lumen volume on quantitative arteriography of a 20-cm segment of the femoral artery. In this report we studied in a representative subgroup (n = 72) whether the reduction in HDL concentrations induced by probucol could explain the failure of the drug to be effective. We analyzed the effects of treatment on HDL particle size subclasses. Probucol lowered the relative level of HDL2b, comprising the largest HDL particles, by 53% and the protein concentration of HDL2b by 67%. The protein reduction in HDL was mainly confined to the apolipoprotein A-I moiety. The change in lumen volume correlated significantly with change in HDL, ie, HDL cholesterol (r = .34, P < .01), HDL2 cholesterol (r = .37, P < .01), HDL2b protein (r = .44, P < .001), and the relative HDL2b value (r = .51, P < .001). The corresponding values for relative HDL2b, distribution calculated on the active (n = 35) and placebo (n = 37) groups separately were also significant (r = .39 and .32, respectively; both P < .05). The correlation between drug-induced change in the relative HDL2b concentration and change in atherosclerosis was independent of the alteration in triglyceride concentration and could not be explained by treatment interaction.(ABSTRACT TRUNCATED AT 250 WORDS)