OUP user menu

Factors underlying regression of coronary atheroma with potent statin therapy

Rishi Puri, Steven E. Nissen, Christie M. Ballantyne, Phillip J. Barter, M. John Chapman, Raimund Erbel, Peter Libby, Joel S. Raichlen, Julie St. John, Kathy Wolski, Kiyoko Uno, Yu Kataoka, Stephen J. Nicholls
DOI: http://dx.doi.org/10.1093/eurheartj/eht084 1818-1825 First published online: 3 May 2013

Abstract

Aims Statins can inhibit the progression of coronary atherosclerosis. We aimed to characterize clinical factors that associate with differing measures of coronary atheroma volume following potent statin therapy.

Methods and results SATURN employed serial intravascular ultrasound (IVUS) to monitor changes in measures of coronary atheroma burden [total atheroma volume (TAV) and per cent atheroma volume (PAV)] in 1039 patients with coronary artery disease, treated with rosuvastatin (40 mg) or atorvastatin (80 mg) daily for 24 months. Rosuvastatin-treated patients demonstrated greater reductions in low-density lipoprotein cholesterol (LDL-C, 47 vs. 40%, P < 0.001) and greater increases in high-density lipoprotein cholesterol (HDL-C, 13 vs. 10%, P = 0.02). These alterations in the lipid profile associated with greater TAV (−6.4 vs. −4.4 mm3, P = 0.01), but not PAV (−1.22 vs. −0.99%, P = 0.17) regression. Greater TAV reductions with rosuvastatin vs. atorvastatin occurred in patients with diabetes (P = 0.01, treatment by diabetic status interaction P-value 0.05). Greater PAV reductions with rosuvastatin were evident in females (P = 0.01, treatment by sex interaction P-value 0.03) and in those with greater than or equal to median baseline LDL-C (P = 0.02, treatment by LDL-C group interaction P-value 0.03) or HDL-C levels (P = 0.02, treatment by HDL-C group interaction P-value 0.04). On multivariable analysis assessing change in TAV and PAV, both higher baseline TAV and PAV independently associated with TAV and PAV regression, respectively (standardized estimates: TAV −0.25, P < 0.001; PAV −0.23, P < 0.001).

Conclusion Higher-risk patients, particularly those with greater baseline coronary atheroma volume, are more likely to experience less disease progression with potent statin therapy.

  • IVUS
  • Statins
  • Atherosclerosis
  • Risk factors

Introduction

In patients at high risk for future cardiovascular events, high-dose statin regimens improve clinical outcomes.14 Accumulating evidence from clinical trials utilizing serial coronary intravascular ultrasound (IVUS) demonstrates that intensive statin therapy can halt the progression of atherosclerosis,5 and sometimes induces disease regression.6 Yet, the factors underlying differential effects of individual statins on disease progression remain poorly characterized.

Preclinical studies have demonstrated that statins exert favourable effects beyond their ability to lower levels of atherogenic lipoproteins.7 In addition to their beneficial impact on oxidation and inflammatory pathways implicated in atherogenesis and plaque rupture,8 statins may also exert a direct effect upon the arterial wall.9 Such actions may be important, given that arterial wall remodelling plays a pivotal role in the clinical expression of atherosclerotic disease. This role is highlighted by observations that culprit lesions in the setting of acute coronary syndromes typically demonstrate expansive arterial remodelling10 and that measures of atheroma volume incorporating the influence of remodelling associate more closely with the incidence of cardiovascular events.11

It is important, therefore, to understand whether factors that influence different measures of progression or regression of coronary atherosclerosis are similar. Of particular interest in the Study of Coronary Atheroma by Intravascular Ultrasound: Effect of Rosuvastatin Versus Atorvastatin (SATURN) was that the highest dose of rosuvastatin produced greater regression of total atheroma volume (TAV), but not of per cent atheroma volume (PAV), when compared with high-dose atorvastatin.12 Accordingly, the current analysis aimed to characterize the factors associated with regression of different measures of coronary atherosclerosis in SATURN.

Methods

Patient selection

The design of SATURN has been previously described.13 Briefly, patients with angiographic coronary artery disease and low-density lipoprotein cholesterol (LDL-C) <116 mg/dL following a 2-week treatment period with atorvastatin (40 mg) or rosuvastatin (20 mg) daily were re-randomized and treated for 24 months with atorvastatin (80 mg) or rosuvastatin (40 mg) daily. Intravascular ultrasound imaging was performed within a coronary artery at baseline and following 104 weeks of treatment.

Acquisition and analysis of intravascular coronary imaging

The acquisition and analysis of IVUS images has been described in detail previously.5,6,11,1418 Patients were required to have coronary artery disease, defined as the presence of at least one lumen stenosis >20% in an epicardial coronary artery at the time of a clinically indicated coronary angiogram. Imaging was performed in a single, native coronary artery with no lumen stenosis >50%, which had not undergone revascularization and was not considered to be the culprit vessel for a prior myocardial infarction. Anatomically matched arterial segments were selected for analysis on the basis of a proximal and distal-side branch (fiduciary points). Cross-sectional images spaced 1-mm apart were selected for analysis, with the lumen and the external elastic membrane leading edges defined by manual planimetry. The plaque area was determined as the area between these leading edges. Per cent atheroma volume and TAV were calculated as previously described.11 Changes in coronary atheroma volume (PAV and TAV) were calculated as the atheroma volume at 104 weeks minus the corresponding volume at baseline. Plaque regression was defined as any decrease in atheroma volume from baseline.

Statistical analysis

Continuous variables were compared between groups using Student's t-test or the Wilcoxon rank-sum test with mean ± SD or median (inter-quartile range; IQR) reported, respectively. Categorical variables were compared between groups using the Pearson χ2 test with frequency and per cent reported. Changes from baseline in biochemical parameters were assessed using the Wilcoxon signed-rank test. Change in TAV and PAV are reported as median (distribution-free 95% confidence interval) with treatment groups compared using the analysis of covariance (ANCOVA) on rank-transformed data, while adjusting for baseline values and geographic region. A multivariable linear regression model was used to determine factors associated with changes in TAV and PAV, and results are expressed as standardized and parameter estimates. Factors for consideration in the model included demographics, medical history, vital signs, prior medications, baseline and on-treatment lipids, baseline IVUS measures, and treatment group. All tests were two-sided with a P-value <0.05 considered significant. All statistical analyses were performed using the SAS software, version 9.2 (SAS Institute, Cary, NC, USA).

Results

Patient characteristics

Of 1385 randomized patients, 1039 had evaluable imaging at baseline and follow-up that enabled comparison of atorvastatin (n = 519) and rosuvastatin (n = 520) treatment effects. No differences between treatment groups were observed with regard to patient demographics or concomitant medication use (Table 1). Similarly, baseline clinical characteristics were not different between patients completing the study and those who did not complete it.12

View this table:
Table 1

Baseline characteristics of patients

CharacteristicTotal (n = 1039)Atorvastatin (n = 519)Rosuvastatin (n = 520)P-value
Age (years)57.6 ± 8.657.9 ± 8.557.4 ± 8.60.32
Male sex765 (73.6)386 (74.4)379 (72.9)0.59
Diabetes159 (15.3)87 (16.8)72 (13.8)0.19
Hypertension731 (70.4)367 (70.7)364 (70.0)0.80
Current smoking336 (32.3)157 (30.3)179 (34.4)0.15
Prior statin usea622 (59.9)319 (61.5)303 (58.3)0.29
Any anti-platelet therapy1015 (97.7)508 (97.9)507 (97.5)0.68
Beta-blocker632 (60.8)317 (61.1)315 (60.6)0.87
ACE inhibitor457 (44.0)231 (44.5)226 (43.5)0.73
Angiotensin receptor blocker (ARB)170 (16.4)82 (15.8)88 (16.9)0.63
  • Values refer to the intention-to-treat population.

  • Mean ± SD or n (%) are reported.

  • Medications are reported as concomitant therapies.

  • ACE, angiotensin-converting enzyme.

  • aPrior statin use was defined as use of a statin within 30 days prior to enrolment in the trial.

Biochemical measurements

Table 2 shows baseline levels and per cent changes of lipid variables and C-reactive protein in both treatment groups. Compared with patients receiving atorvastatin, the rosuvastatin-treated group experienced greater reductions in total cholesterol (−27.7 ± 15.1 vs. −24.8 ± 16.0%, P = 0.004), LDL-C (−47.2 ± 18.8 vs. −40.1 ± 19.9%, P < 0.001), and apolipoprotein B (apoB, −30.4 ± 16.3 vs. −26.9 ± 19.7%, P < 0.001), and greater increases in high-density lipoprotein cholesterol (HDL-C, 12.7 ± 18.6 vs. 10.0 ± 17.94%, P = 0.02) and apolipoprotein A-I (apoA-I, 16.2 ± 17.6 vs. 10.6 ± 17.7%, P < 0.001). The atorvastatin-treated group experienced greater per cent reductions in triglyceride levels [−13.2% (−32.8, 14.7) vs. −6.2% (−29.2, 20.8), P = 0.007]. C-reactive protein decreased by 25.0% (−55.9, 21.3) with atorvastatin and 23.7% (−57.1, 44.1) with rosuvastatin (P = 0.34).

View this table:
Table 2

Baseline and per cent changes of biochemical values during treatment

VariableAt baselineP-value% change during treatmentP-value
AtorvastatinRosuvastatinAtorvastatinRosuvastatin
Cholesterol
 Total (mg/dL)193.5 ± 34.2193.9 ± 34.10.86−24.8 ± 16.0−27.7 ± 15.10.004
 LDL-C (mg/dL)119.9 ± 28.9120.0 ± 27.30.94−40.1 ± 19.9−47.2 ± 18.8<0.001
 HDL-C (mg/dL)44.7 ± 10.745.3 ± 11.80.4110.0 ± 17.912.7 ± 18.60.02
 LDL-C:HDL-C2.8 ± 0.92.8 ± 0.90.81−44.7 ± 18.9−52.1 ± 17.5<0.001
Triglycerides (mg/dL)
 Median1301280.55−13.2−6.20.007
 Inter-quartile range97, 17791, 181−32.8, 14.7−29.2, 20.8
Apolipoprotein
 B (mg/dL)104.9 ± 21.7105.4 ± 21.20.68−26.9 ± 19.7−30.4 ± 16.30.002
 A-1 (mg/dL)126.2 ± 23.3128.0 ± 25.20.2310.6 ± 17.716.2 ± 17.6<0.001
C-reactive proteina
 Median1.51.70.29−25.0−23.70.34
 Inter-quartile range0.8, 3.30.8, 3.8−55.9, 21.3−57.1, 44.1
  • Values refer to the intention-to-treat population.

  • Mean ± SD or median (IQR) are reported.

  • Medians and inter-quartile ranges for variables that were not normally distributed were calculated with the use of the Wilcoxon rank-sum test.

  • LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol.

  • aFinal measurements were used for values obtained during treatment.

Changes in plaque volume: subgroup analysis

Baseline TAV and PAV did not differ significantly between the rosuvastatin and atorvastatin groups (TAV: 144.1 ± 61 vs. 144.2 ± 64 mm3, P = 0.99; PAV: 36.7 ± 8.2 vs. 36.0 ± 8.3%, P = 0.33). On serial evaluation, we noted differences in treatment effects between the two statin regimens were noted. The rosuvastatin group experienced a greater reduction than atorvastatin-treated patients in TAV [−6.4 (95% CI: −7.52, −5.12) vs. −4.4 mm3 (95% CI: −5.98, −3.26), P = 0.01], but not in PAV [−1.22 (95% CI: −1.52, −0.90) vs. −0.99% (95% CI: −1.19, −0.63), P = 0.17]. Compared with the atorvastatin group, a greater percentage of rosuvastatin-treated patients demonstrated regression of TAV (71.3 vs. 64.7%, P = 0.02), but not of PAV (68.5 vs. 63.2%, P = 0.07).

Tables 3 and 4 summarize subgroups that underwent changes in TAV and PAV and the comparative effects between statins. Significantly greater TAV regression occurred in those patients with greater than median values for baseline TAV (−11.1 vs. −8.1 mm3, P = 0.01), age (−6.0 vs. −4.1 mm3, P = 0.03), systolic blood pressure (−6.8 vs. −4.7 mm3, P = 0.03), baseline LDL-C (−7.0 vs. −4.6 mm3, P = 0.03), and triglycerides (−5.7 vs. −4.3 mm3, P = 0.045), as well as women (−7.2 vs. −3.2 mm3, P = 0.03), patients with diabetes (−8.6 vs. −0.9 mm3, P = 0.01), and in patients who were statin naive in the month prior to randomization (−7.2 vs. −3.4 mm3, P = 0.03). The only significant interaction between treatment groups, however, occurred in diabetic vs. non-diabetic patients (P = 0.05). This suggests that rosuvastatin had greater TAV-lowering effects vs. atorvastatin in patients with diabetes mellitus, than in patients without the disease.

View this table:
Table 3

Subgroup analysis of predictors of changes in total atheroma volume

Baseline characteristic n, n (atorvastatin, rosuvastatin)Atorvastatin median (95% CI)P-value for change from baselineRosuvastatin median (95% CI)P-value for change from baselineP-value for treatment between groupsP-value for interaction
Age (years) <median (252, 253)−5.22 (−7.57, −3.24)<0.001−6.42 (−9.02, −4.93)<0.0010.140.65
Age (years) ≥median (267, 267)−4.09 (−5.98, −2.52)<0.001−5.97 (−8.10, −4.37)<0.0010.03
Men (386, 379)−5.15 (−7.14, −3.56)<0.001−5.88 (−7.44, −4.89)<0.0010.090.33
Women (133, 141)−3.17 (−5.70, −0.36)<0.001−7.15 (−9.70, −4.27)<0.0010.03
Baseline SBP (mmHg) <median (230, 241)−4.24 (−6.57, −1.59)<0.001−5.72 (−7.50, −4.27)<0.0010.130.75
Baseline SBP (mmHg) ≥median (289, 279)−4.65 (−6.89, −3.30)<0.001−6.77 (−8.82, −5.19)<0.0010.03
Non-diabetic (432, 448)−5.08 (−6.62, −3.53)<0.001−6.10 (−7.23, −4.79)<0.0010.110.05
Diabetic (87, 72)−0.86 (−5.78, 2.61)0.10−8.61 (−13.4, −4.95)<0.0010.01
No statin use 30 days pre-consent (200, 217)−3.43 (−6.39, −2.45)<0.001−7.23 (−9.25, −5.24)<0.0010.030.48
Statin use 30 days pre-consent (319, 303)−5.05 (−6.63, −3.51)<0.001−5.36 (−7.28, −4.37)<0.0010.13
Baseline LDL-C (mg/dL) <median (254, 252)−4.46 (−7.51, −2.47)<0.001−5.33 (−7.3, −4.01)<0.0010.240.51
Baseline LDL-C (mg/dL) ≥median (262, 264)−4.61 (−6.37, −3.17)<0.001−7.02 (−8.82, −5.18)<0.0010.03
Baseline HDL-C (mg/dL) <median (252, 258)−4.84 (−7.60, −2.50)<0.001−6.17 (−7.88, −4.59)<0.0010.060.86
Baseline HDL-C (mg/dL) ≥median (266, 259)−4.32 (−5.77, −3.06)<0.001−6.53 (−8.82, −4.40)<0.0010.10
Baseline triglycerides (mg/dL) <median (253, 260)−5.05 (−7.35, −3.29)<0.001−6.45 (−8.78, −4.94)<0.0010.150.70
Baseline triglycerides (mg/dL) ≥median (265, 257)−4.27 (−6.04, −2.50)<0.001−5.70 (−8.01, −4.27)<0.0010.045
Baseline TAV (mm3) <median (254, 266)−2.72 (−4.21, −0.76)<0.001−4.05 (−5.25, −2.64)<0.0010.250.30
Baseline TAV (mm3) ≥median (265, 254)−8.14 (−10.41, −5.21)<0.001−11.07 (−14.71, −8.54)<0.0010.01
  • SBP, systolic blood pressure;  LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol.

View this table:
Table 4

Subgroup analysis of predictors of changes in per cent atheroma volume

Baseline Characteristic n,n (atorvastatin, rosuvastatin)Atorvastatin Median (95% CI)P-value for change from baselineRosuvastatin Median (95% CI)P-value for change from baselineP-value for treatment between groupsP-value for interaction
Age (years) <median (252, 253)−1.06 (−1.43, −0.46)<0.001−1.12 (−1.60, −0.73)<0.0010.960.22
Age (years) ≥median (267, 267)−0.84 (−1.28, −0.46)<0.001−1.35 (−1.68, −0.90)<0.0010.07
Men (386, 379)−1.03 (−1.32, −0.70)<0.001−1.09 (−1.44, −0.72)<0.0010.990.03
Women (133, 141)−0.71 (−1.38, −0.25)<0.001−1.76 (−2.39, −1.02)<0.0010.01
Baseline SBP (mmHg) <median (230, 241)−1.08 (−1.44, −0.40)<0.001−1.14 (−1.80, −0.69)<0.0010.110.34
Baseline SBP (mmHg) ≥median (289, 279)−0.84 (−1.28, −0.49)<0.001−1.30 (−1.55, −0.85)<0.0010.74
Non-diabetic (432, 448)−1.04 (−1.35, −0.70)<0.001−1.31 (−1.53, −0.91)<0.0010.180.69
Diabetic (87, 72)−0.50 (−1.30, 0.01)0.001−0.86 (−1.86, −0.31)<0.0010.89
No statin use 30 days pre-consent (200, 217)−0.84 (−1.30, −0.39)<0.001−1.37 (−1.83, −0.85)<0.0010.100.30
Statin use 30 days pre-consent (319, 303)−1.01 (−1.41, −0.50)<0.001−1.14 (−1.55, −0.70)<0.0010.72
Baseline LDL-C (mg/dL) <median (254, 252)−1.00 (−1.45, −0.50)<0.001−0.95 (−1.52, −0.50)<0.0010.500.03
Baseline LDL-C (mg/dL) ≥median (262, 264)−1.00 (−1.28, −0.41)<0.001−1.44 (−1.77, −1.02)<0.0010.02
Baseline HDL-C (mg/dL) <median (252, 258)−1.20 (−1.73, −0.72)<0.001−1.13 (−1.55, −0.69)<0.0010.560.04
Baseline HDL-C (mg/dL) ≥median (266, 259)−0.68 (−1.04, −0.25)<0.001−1.35 (−1.67, −0.85)<0.0010.02
Baseline triglycerides (mg/dL) <median (253, 260)−0.84 (−1.32, −0.44)<0.001−1.26 (−1.57, −0.71)<0.0010.360.99
Baseline triglycerides (mg/dL) ≥median (265, 257)−1.04 (−1.38, −0.54)<0.001−1.14 (−1.76, −0.82)<0.0010.37
Baseline PAV (%) <median (259, 260)−0.22 (−0.54, 0.12)0.04−0.69 (−1.05, −0.24)<0.0010.040.16
Baseline PAV (%) ≥median (260, 260)−1.89 (−2.15, −1.19)<0.001−1.72 (−2.07, −1.41)<0.0010.96
  • SBP, systolic blood pressure; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol.

Greater reductions of PAV were found in rosuvastatin-treated patients vs. atorvastatin-treated patients in those with less than median baseline PAV values (−0.69 vs.−0.22%, P = 0.04), in women (−1.76 vs. −0.71%, P = 0.01) and in patients with greater than median baseline LDL-C (−1.44 vs. −1.00%, P = 0.02) or HDL-C levels (−1.35 vs. −0.68%, P = 0.02). There were significant interactions between treatment and sex (P = 0.03), baseline LDL-C group (P = 0.03) and baseline HDL-C group (P = 0.04). This suggests that rosuvastatin vs. atorvastatin had greater PAV-lowering effects in women than in men, and in those with higher baseline LDL-C or HDL-C levels than in those with lower levels.

Multivariable models for changes in coronary atheroma volume

We performed a multivariable linear regression analysis to identify systematically factors that associated with changes in TAV and PAV (Tables 5 and 6). The standardized estimates show that in comparison with other variables in the model, increasing baseline TAV predicted TAV regression (standardized estimate −0.25, P < 0.001), as did randomization to rosuvastatin (standardized estimate −0.071, P = 0.019). Average on-treatment non-HDL cholesterol levels, older age, increasing weight at baseline, prior percutaneous coronary intervention, and prior beta-blocker therapy were each independent predictors of TAV progression (Table 5).

View this table:
Table 5

Multivariable model for predicting changes in total atheroma volume

VariableStandardized estimateParameter estimateP-value
Baseline TAV (mm3)−0.25−0.061<0.001
Non-HDL cholesterol (mg/dL)a0.110.064<0.001
Age (at baseline) (years)0.0920.160.003
Weight (at baseline) (kg)0.0730.0620.02
Rosuvastatin treatment group−0.071−2.120.02
Prior coronary angioplasty or stent0.0662.330.03
Prior beta-blockers0.0632.080.04
  • aAverage on-treatment value.

View this table:
Table 6

Multivariable model for predicting changes in per cent atheroma volume

VariableStandardized estimateParameter estimateP-value
Baseline PAV (%)−0.23−0.082<0.001
Prior nitrates0.0910.920.003
Female sex−0.081−0.540.008
LDL-C (mg/dL)a0.0780.0100.009
  • aAverage on-treatment value.

Similarly, increasing baseline PAV predicted PAV regression (standardized estimate −0.23, P < 0.001). Female sex was also associated with PAV regression (standardized estimate −0.081, P = 0.008). Prior nitrate therapy and increasing average on-treatment LDL-C level each associated with PAV progression (Table 6).

Discussion

Statins have demonstrated efficacy in the management of atherosclerotic disease, and high-dose statin therapy associates with lower clinical event rates,1 as well as with the attenuation of plaque progression.5 SATURN was the first head-to-head comparison of the two most efficacious statins currently available, each prescribed at their highest approved dosages. Both treatments produced net plaque regression during a 2-year period. Although we did not see a net difference in the primary efficacy endpoint (change in PAV) between each treatment group, rosuvastatin produced greater reductions of TAV. Differing clinical and biochemical factors associated with changes in TAV and PAV, with some differences noted between treatment groups. In the rosuvastatin-treated group vs. the atorvastatin-treated group, TAV regression was more likely to occur in patients with diabetes mellitus, whereas PAV regression was more evident in women and in those with higher baseline LDL-C or HDL-C levels. These differences highlight the variable influence of clinical and biochemical factors upon atherosclerotic plaque and arterial remodelling, in response to intensive statin treatment. On multivariable analysis for predicting changes in TAV and PAV, we consistently found that a greater baseline coronary atheroma volume (measured as both TAV and PAV) associated with more disease regression with intensive statin treatment.

The results of this analysis have further implications for the management of patients at risk for future cardiovascular events. Although increasingly utilized as an imaging biomarker for evaluating the mechanistic efficacy of novel anti-atherosclerotic agents,5,6,1421 the burden of coronary atherosclerosis is increasingly reported to predict future cardiovascular events.11,2225 While IVUS is the ‘gold-standard’ for quantifying coronary atheroma volume, emerging data from non-invasive coronary imaging studies corroborate the link between coronary atheroma volume and clinical events.26,27 Atherosclerosis reflects a lifetime of cardiovascular risk factors, and predisposes to coronary events. But current risk-prediction algorithms have limited prognostic ability,28,29 and volumetric measures of coronary plaque correlate modestly with cardiovascular risk factors.30 So should coronary plaque burden be incorporated into clinical practice as a measure of coronary risk? If accurately measured with a non-invasive technique, the emerging data would support this conjecture. This hypothesis would, however, require extensive prospective validation. Furthermore, the results of this analysis would suggest that those at greatest coronary risk, who notionally harbour the greatest burden of disease, would stand to benefit the most from intensive statin treatment, by experiencing greater degrees of plaque regression.

Women randomized to the rosuvastatin group achieved significantly more PAV regression than those randomized to atorvastatin, a difference that was greater than in men. On multivariable analysis, female sex was an independent predictor of PAV regression. Opinions are mixed regarding the benefits of statin treatment in women,31,32 resulting in differential prescribing patterns of statins in men and women for the management of coronary artery disease,33,34 despite equally high mortality rates in both sexes following acute coronary syndromes. But a recent post hoc analysis of the Pravastatin or Atorvastatin Evaluation and Infection Therapy – Thrombolysis in Myocardial Infarction 22 (PROVE IT-TIMI 22) trial found that potent statin therapy (atorvastatin, 80 mg) in women following acute coronary syndromes resulted in marked reductions in clinical events compared with standard lipid lowering (pravastatin, 40 mg).35 This result occurred despite women experiencing less LDL-C lowering with intensive statin treatment than men. Our findings of greater PAV regression in women provide mechanistic support for the observations made in PROVE IT-TIMI 22, confirming the greater benefit of intensive statin treatment in at-risk women. The underlying biological mechanisms of these findings warrant further investigation.

Increasing age, on multivariable analysis, associated with more progression of TAV. Older patients harbour more progressive, calcified atherosclerosis, independent of other coronary risk factors.36 Yet patients achieving LDL-C levels <80 mg/dL still had consistently slower rates of disease progression. As a result, atheroma regression in older individuals in general may be more difficult to achieve. This may be a result of more calcified disease observed in older individuals. Conversely, these results from SATURN also raise the question of whether disease regression occurs more readily in younger individuals, who may harbour more malleable coronary atheroma. These observations support the notion of interventions that aggressively target cardiovascular risk factors, particularly in younger individuals, who may stand to derive the greatest clinical benefit during their lifespan. Furthermore, given that age impacted the change in TAV, but not PAV, changes in atheroma volume in older patients may less likely involve corresponding changes to arterial wall remodelling, which is reflected in the calculation of PAV.

Higher average on-treatment LDL-C and non-HDL cholesterol levels were associated on multivariable analysis with PAV and TAV progression. Consistent with prior studies that assessed the volume of coronary atherosclerosis by PAV, TAV data from the same IVUS-based imaging trials highlight a similar linear relationship between the degree of coronary atheroma regression achieved and the on-treatment LDL-C level (Figure 1). Although mean on-treatment LDL-C levels in SATURN were among the lowest achieved of any previously conducted atherosclerosis imaging study (62 mg/dL in rosuvastatin-treated patients and 70 mg/dL in atorvastatin-treated patients), one-third of patients still experienced disease progression. This observation has implications for the development of novel therapeutic strategies aimed at lowering the burden of atherosclerotic disease, and for addressing residual event rates observed in patients prescribed contemporary anti-atherosclerotic therapies. While the quest for novel therapeutic targets continues, continuing to lower LDL-C levels to the order of 30–40 mg/dL may achieve even further degrees of atheroma regression, as suggested by Figure 1. A combined therapeutic approach may meet this need of further and/or alternative avenues of LDL-C reduction. This conjecture, however, will require testing in a randomized trial.

Figure 1

Median changes in total atheroma volume (TAV) and per cent atheroma volume (PAV) vs. average on-treatment low-density lipoprotein cholesterol levels in imaging trials, using serial intravascular ultrasound to evaluate various anti-atherosclerotic strategies. SATURN: The Study of Coronary Atheroma by Intravascular Ultrasound: Effect of Rosuvastatin vs. Atorvastatin; ASTEROID: A Study to Evaluate the Effect of Rosuvastatin on Intravascular Ultrasound-Derived Coronary Atheroma Burden; ILLUSTRATE: Investigation of Lipid Level Management Using Coronary Ultrasound To Assess Reduction of Atherosclerosis by CETP Inhibition and HDL Elevation; REVERSAL: Reversal of Atherosclerosis with Aggressive Lipid Lowering; STRADIVARIUS: Strategy To Reduce Atherosclerosis Development InVolving Administration of Rimonabant—the Intravascular Ultrasound Study; CAMELOT: Comparison of Amlodipine vs. Enalapril to Limit Occurrences of Thrombosis.

Beta-blocker therapy, on multivariable analysis, associated with TAV progression. On the contrary, a pooled analysis of serial changes of IVUS-derived TAV from four clinical trials concluded that beta-blocker therapy slowed the progression of coronary atherosclerosis.37 This analysis was, however, limited by its non-randomized design and unequal distribution of different beta-blockers, and on-treatment LDL-C levels were above those recommended by current treatment guidelines.38 While our findings suggest that beta-blocker use is not associated with the progression of coronary atherosclerosis in the setting of very low (<70 mg/dL) LDL-C levels, we cannot exclude the possibility of confounding whereby beta-blockers were administered to higher-risk patients, with more progressive disease, in SATURN. Indeed, prior investigators have found potential pro-atherogenic effects of propranolol in patients with angina pectoris,39,40 and the use of beta-blocker therapy in such patients remains controversial.41

Rosuvastatin and atorvastatin produced similar, significant reductions in PAV.12 But treatment with rosuvastatin caused significantly greater reduction of TAV, and regression occurred in a greater proportion of rosuvastatin-treated patients than in corresponding atorvastatin-treated patients. Although PAV and TAV are well-validated measures of IVUS-derived plaque volume,42,43 PAV incorporates the amount of plaque present in relation to the adaptive response of the vessel wall. As such, the concomitant arterial remodelling response of the vessel wall affects the calculation of PAV. Arterial remodelling influences the clinical response of atherosclerotic plaque. Lesions in patients with unstable coronary syndromes were found to harbour significantly greater atheroma volume and more frequent expansive arterial remodelling than lesions in patients with stable coronary syndromes.10 With its association with clinical events,11 smaller coefficient of variation, and subsequent impact on smaller sample sizes tested in clinical trials, PAV preferentially evolved as the primary efficacy endpoint of most trials that utilize serial IVUS to assess changes in coronary atheroma volume. Limited data exist, however, regarding the influence of lipid-lowering therapies on coronary arterial remodelling. In a post hoc analysis of the Reversal of Atherosclerosis With Aggressive Lipid Lowering (REVERSAL) trial, constrictive arterial remodelling was the predominant response during high-dose atorvastatin therapy, which associated with the anti-inflammatory effects of this treatment.9 Our analysis would suggest that sex differences may also significantly influence the nature of plaque progression and remodelling (measured as PAV), along with achieved LDL-C levels, in the setting of intensive statin therapy.

Some caveats of the current analysis should be noted. Approximately 25% of patients who were initially enrolled in SATURN failed to undergo a follow-up IVUS. These patients may have experienced rates of plaque progression or regression that could have influenced the results presented in this analysis. SATURN was not a placebo-controlled study, and so it was not possible to directly compare these results to the natural history of atherosclerosis in such patients. This trial enrolled patients who had symptoms requiring coronary angiography. As such, the study may not apply to primary prevention. Moreover, the direct link between reduction in coronary atheroma volume assessed by IVUS and clinical event prevention requires further study. Given that SATURN is the largest and only head-to-head imaging trial comparing these two potent statins, a study of significant magnitude with prolonged follow-up would be required to ascertain meaningful clinical differences between treatment regimens. This type of study is unlikely to ever be conducted.

Conclusions

Statins are the cornerstone of pharmacological management of atherosclerotic disease. SATURN highlighted the marked anti-atherosclerotic effects of two potent statin regimens, each resulting in significant coronary atheroma volume regression. In the setting of potent statin therapy, higher-risk patients, particularly those with higher baseline coronary atheroma volume, consistently experienced disease regression. This was irrespective of whether atheroma volume was measured as TAV or PAV. Female sex and achieved LDL-C levels associated with regression and progression of PAV, respectively—a measure of plaque burden that also incorporates the arterial remodelling response. Despite very low on-treatment LDL-C levels, nearly one-third of patients still demonstrated disease progression. These findings outline the variable clinical and biochemical characteristics that affect the natural course of coronary atherosclerosis, the arterial wall response, and highlight the need for additional therapeutic targets to tackle the residual burden of disease.

Funding

Dr Puri is jointly supported by a Postgraduate Medical Research Scholarship from the National Health & Medical Research Council (565579), the National Heart Foundation of Australia (PC0804045) and Dawes Scholarships (Hanson Institute). SATURN was funded by AstraZeneca Pharmaceuticals.

Conflict of interest: Dr Nissen has received research support to perform clinical trials through the Cleveland Clinic Coordinating Center for Clinical Research from Pfizer, AstraZeneca, Novartis, Roche, Daiichi-Sankyo, Takeda, Sanofi-Aventis, Resverlogix and Eli Lilly; and is a consultant/advisor for many pharmaceutical companies but requires them to donate all honoraria or consulting fees directly to charity so that he receives neither income nor a tax deduction. Dr Nicholls has received speaking honoraria from AstraZeneca, Pfizer, Merck Schering-Plough and Takeda; consulting fees from AstraZeneca, Pfizer, Merck Schering-Plough, Takeda, Roche, NovoNordisk, LipoScience, and Anthera; research support from AstraZeneca and Lipid Sciences. Dr Ballantyne has received grant support from Abbott, Astra-Zeneca, Bristol-Myers Squibb, Genentech, GlaxoSmithKline, Kowa, Merck, Novartis, Roche, Sanofi-Synthelabo and Takeda; consulting fees and honoraria from Abbott, Adnexus, Amarin, Amylin, AstraZeneca, Bristol-Myers Squibb, Esperion, Genentech, GlaxoSmithKline, Idera, Kowa, Merck, Novartis, Omthera, Resverlogix, Roche, Sanofi-Synthelabo, and Takeda; lecture fees from Abbott, AstraZeneca, GlaxoSmithKline and Merck. Dr Barter holds an advisory board position for AstraZeneca, Merck, Roche, CSL, Behring and Pfizer; receives grant support from Merck; consulting fees from CSL Behring; lecture fees from AstraZeneca, Kowa, Merck, Pfizer and Roche. Dr Chapman receives grant support from Merck and Kowa; consulting fees from Merck and Pfizer; lectures fees from Merck and Kowa. Dr Erbel receives grant and travel support from the Heinz Nixdorf Foundation, German Research Foundation; accommodations/meeting expenses from Biotronik, Sanofi, and Novartis. Dr Libby serves as an unpaid consultant for Novartis, Johnson & Johnson, Amgen, and Roche; serves in unpaid leadership roles for clinical trials sponsored by AstraZeneca, GlaxoSmithKline, Novartis, Pfizer, Pronova, and Sigma Tau; and previously received royalties from Roche for the patent on CD40L in cardiovascular risk stratification. Dr Raichlen is an employee and owns stock in AstraZeneca. All other authors have no disclosures.

References

View Abstract