High-intensity statin treatment is associated with reduced plaque structural stress and remodelling of artery geometry and plaque architecture

Abstract Aims Plaque structural stress (PSS) is a major cause of atherosclerotic plaque rupture and major adverse cardiovascular events (MACE). We examined the predictors of changes in peak and mean PSS (ΔPSSpeak, ΔPSSmean) in three studies of patients receiving either standard medical or high-intensity statin (HIS) treatment. Methods and results We examined changes in PSS, plaque size, and composition between 7348 co-registered baseline and follow-up virtual-histology intravascular ultrasound images in patients receiving standard medical treatment (controls, n = 18) or HIS (atorvastatin 80 mg, n = 20, or rosuvastatin 40 mg, n = 22). The relationship between changes in PSSpeak and plaque burden (PB) differed significantly between HIS and control groups (P < 0.001). Notably, PSSpeak increased significantly in control lesions with PB >60% (P = 0.04), but not with HIS treatment. However, ΔPSSpeak correlated poorly with changes in lumen and plaque area or PB, plaque composition, or lipid lowering. In contrast, ΔPSSpeak correlated significantly with changes in lumen curvature, irregularity, and roughness (P < 0.05), all of which were reduced in HIS patients. ΔPSSmean correlated with changes in lumen area, PA, PB, and circumferential calcification, and was unchanged with either treatment. Conclusion Our observational study shows that PSSpeak changes over time were associated with baseline disease severity and treatment. The PSSpeak increase seen in advanced lesions with standard treatment was associated with remodelling artery geometry and plaque architecture, but this was not seen after HIS treatment. Smoothing plaques by reducing plaque/lumen roughness, irregularity, and curvature represents a novel mechanism whereby HIS may reduce PSS and, thus may protect against plaque rupture and MACE.

where [ @ ] and [ @D ] are the displacement vector and stress tensor, respectively, is the density of each component, and is time.
The entire plaque geometric model was meshed using 9-node quadrilaterals, generating approximately 10,000 elements and 40,000 nodes per model. Displacement and strain were assumed to be large. There was no relative movement at the interface of atherosclerotic components and the relative energy tolerance was set to be 0.005. Two adjacent points located on the outer wall were fixed to prevent rigid body displacement. Maximum principal stress was used to characterize the mechanical loading within the plaque structure (PSS) in the periluminal region (0.2mm maximum depth from the luminal contour). Mean PSS was calculated as the mean value of PSS experienced by all the luminal nodes. Dynamic loading conditions were standardized to 120/70mmHg. Pressure at the outer boundary was set to zero. All simulations were performed using ADINA 9.5 (ADINA R&D, Inc., USA) software. Figure S1):

Additional measures (
• Lumen aspect ratio = maximum diameter of ellipse (or lumen major axis)/minimum diameter of ellipse (or lumen minor axis), i.e., lower (improved) aspect ratio describes a rounder lumen, and a value of 1 indicates a perfectly circular lumen.
• Lumen curvature: 3,4 curvature at point a (in Figure S1) was computed using the radius (as ra) of the circle determined by point a and two adjacent points (bottom right figure) on both sides, i.e. curvature = 1/ra. Curvature value was computed for all points in the lumen, and the maximum lumen curvature value (Lumen Curvaturemax) is used in data analysis. The minimum lumen curvature value (Lumen Curvaturemin) is also computed for lumen irregularity calculation • Lumen irregularity 5 = Lumen Curvaturemax -Lumen Curvaturemin • Lumen roughness: reflects the lumen surface evenness in respect to curvature, and calculated using the following formula, with smaller values representing more round or even surface and a perfect round lumen shape will have roughness being 1. Method adapted from. 6 ∆ (r is the radius of the circle best fitting the lumen contour; ra is defined as above in lumen curvature calculation; and ∆l is the length between point a and one adjacent point.)

Assessment of analyst variability
The reproducibility of matching between baseline and follow-up VH-IVUS frames by 2 analysts was examined in 6 vessels that had both baseline (n= 573 frames) and follow-up (n= 623 frames). The 2 analysts reviewed the VH-IVUS data and separately identified the location of follow-up frames in the 2mm segments defined in the baseline frames. To report the intraobserver variability the 1 st analyst performed the analysis twice. The κ test of concordance was used to assess agreement. A good overall agreement was noted for the estimation of the two analysts with the intra-observer variability being 0.733 and the inter-observer variability being 0.701. The reproducibility of lumen curvature, irregularity, and roughness assessment was examined on 2 randomly selected vessels (77 frames) by testing the intraclass correlation coefficient (ICC); this achieved good to excellent absolute agreement: lumen curvature, ICC= 0.787; lumen irregularity, ICC = 0.72; lumen roughness, ICC= 0.712.

Statistical analysis of patient demographics
Continuous variables are presented as mean ± standard deviation or median (interquartile range) and discrete variables as absolute numbers (percentage