Author + information
- Received March 4, 2013
- Revision received March 27, 2013
- Accepted April 9, 2013
- Published online September 3, 2013.
- Ahmed Tawakol, MD∗,†,‡∗ (, )
- Zahi A. Fayad, PhD§,‖,¶,
- Robin Mogg, PhD#,
- Achilles Alon, PharmD#,
- Michael T. Klimas, PhD#,
- Hayes Dansky, MD#,
- Sharath S. Subramanian, MD†,‡,
- Amr Abdelbaky, MD†,‡,
- James H.F. Rudd, MD, PhD∗∗,
- Michael E. Farkouh, MD, MSc¶,††,
- Irene O. Nunes, PhD#,
- Chan R. Beals, MD# and
- Sudha S. Shankar, MD#
- ∗Cardiology Division, Massachusetts General Hospital, Boston, Massachusetts
- †Department of Imaging, Massachusetts General Hospital, Boston, Massachusetts
- ‡Harvard Medical School, Boston, Massachusetts
- §Translational and Molecular Imaging Institute, Mount Sinai School of Medicine, New York, New York
- ‖Department of Radiology, Mount Sinai School of Medicine, New York, New York
- ¶Cardiovascular Institute, Mount Sinai School of Medicine, New York, New York
- #Merck Sharp & Dohme Corporation, Whitehouse Station, New Jersey
- ∗∗Division of Cardiovascular Medicine, University of Cambridge, Cambridge, United Kingdom
- ††Peter Munk Cardiac Centre and Li Ka Shing Knowledge Institute, University of Toronto, Toronto, Ontario, Canada
- ↵∗Reprint requests and correspondence:
Dr. Ahmed Tawakol, MR-PET/CT Program, Massachusetts General Hospital, 165 Cambridge Street, Suite 400, Boston, Massachusetts 02114.
Objectives The study sought to test whether high-dose statin treatment would result in greater reductions in plaque inflammation than low-dose statins, using fluorodeoxyglucose-positron emission tomography/computed tomographic imaging (FDG-PET/CT).
Background Intensification of statin therapy reduces major cardiovascular events.
Methods Adults with risk factors or with established atherosclerosis, who were not taking high-dose statins (n = 83), were randomized to atorvastatin 10 versus 80 mg in a double-blind, multicenter trial. FDG-PET/CT imaging of the ascending thoracic aorta and carotid arteries was performed at baseline, 4, and 12 weeks after randomization and target-to-background ratio (TBR) of FDG uptake within the artery wall was assessed while blinded to time points and treatment.
Results Sixty-seven subjects completed the study, providing imaging data for analysis. At 12 weeks, inflammation (TBR) in the index vessel was significantly reduced from baseline with atorvastatin 80 mg (% reduction [95% confidence interval]: 14.42% [8.7% to 19.8%]; p < 0.001), but not atorvastatin 10 mg (% reduction: 4.2% [–2.3% to 10.4%]; p > 0.1). Atorvastatin 80 mg resulted in significant additional relative reductions in TBR versus atorvastatin 10 mg (10.6% [2.2% to 18.3%]; p = 0.01) at week 12. Reductions from baseline in TBR were seen as early as 4 weeks after randomization with atorvastatin 10 mg (6.4% reduction, p < 0.05) and 80 mg (12.5% reduction, p < 0.001). Changes in TBR did not correlate with lipid profile changes.
Conclusions Statin therapy produced significant rapid dose-dependent reductions in FDG uptake that may represent changes in atherosclerotic plaque inflammation. FDG-PET imaging may be useful in detecting early treatment effects in patients at risk or with established atherosclerosis. (Evaluate the Utility of 18FDG-PET as a Tool to Quantify Atherosclerotic Plaque; NCT00703261)
Inflammation plays an important role in all phases of atherosclerosis, from initiation to atherothrombosis (1,2). Treatment with hydroxymethyl glutaryl coenzyme A (HMG-CoA) reductase inhibitors (statins) results in a marked reduction in atherothrombotic risk, accompanied by substantial reductions in circulating markers of inflammation. However, the effect of statins on atherosclerotic plaque inflammation is less well established.
Positron emission tomography-computed tomographic imaging (PET/CT) with 2-18F-fluoro-2-deoxy-D-glucose (FDG) has been explored for measuring arterial inflammation. FDG accumulates in human atherosclerotic arteries in proportion to macrophage concentration and correlates with gene expression of macrophage-specific markers (3–7). FDG uptake reflects tissue glycolysis, and it has been shown that activated macrophages, especially those stimulation through classical/innate activation pathways (8), have a markedly elevated glycolytic rate (9,10), and thus avidly accumulate FDG (11).
Arterial FDG uptake correlates with the burden of cardiovascular risk factors (12–15), is elevated after recent atherothrombotic events (3,16), and may predict future atherothrombotic risk (17,18). In addition, studies have shown that the vascular FDG PET signal is reproducible (19), allowing serial noninvasive evaluation of plaque inflammation for testing of therapies designed to lower it. Among such therapies, statins are attractive, given their well-described effects on cardiovascular events. Tahara et al. (20) observed, in a single-center open-label study, a reduction in vascular FDG uptake in patients treated with low-dose simvastatin at 12 weeks, compared with patients given dietary advice only. However several important questions remain unanswered by that study: 1) whether the observed statin effect on the arterial PET measurement would be observed under conditions of a double-blind, multicenter trial; 2) whether a dose response effect exists between statin therapy and vascular inflammation; and 3) whether the potential reductions in vascular inflammation can be measured as early as 4 weeks after randomization, given that clinical endpoint trials have demonstrated reductions in cardiovascular endpoints are observed as early as 1 month after initiation of statin therapy (21,22).
Clinical trials have repeatedly demonstrated a greater reduction in cardiovascular events in patients on high- versus low-dose statin therapy (22,23). We hypothesized that high-dose statin therapy would be associated with a greater reduction in plaque inflammation (as measured by FDG uptake) compared with low-dose statin therapy. Accordingly, this represents the first double-blind, multicenter trial using FDG-PET/CT imaging to evaluate the effects of a statin on vascular inflammation.
This double-blind, randomized, active-comparator study was conducted between August 2008 and December 2009 at 10 U.S. centers located near 1 of 6 imaging centers in compliance with the principles of the Declaration of Helsinki and according to Good Clinical Practice guidelines. The protocol was reviewed and approved by each center's institutional review board. All participants provided written informed consent prior to any study procedures.
We sought to study individuals with arterial inflammation. To identify such individuals, we employed a 2-step approach: 1) to identify, via primary screening, individuals at higher risk for atherosclerotic inflammation; and then 2) via secondary screening, confirm the presence of arterial inflammation using PET/CT imaging.
Accordingly, men and women age 30 to 80 years were included if they had documentation or history of any 1 of the following: 1) coronary artery disease; 2) carotid artery disease; 3) cerebrovascular disease; 4) peripheral arterial disease defined by an ankle-brachial index ≥0.5 and ≤0.9; 5) type 2 diabetes mellitus; or 6) body mass index 30 to 40 kg/m2 (inclusive) and waist circumference >102 cm in men, and >88 cm in women. Additional inclusion criteria were low-density lipoprotein (LDL) cholesterol ≥60 mg/dl and triglyceride level <350 mg/dl. Patients were either statin naive or on low dose of a statin, defined as atorvastatin ≤10 mg, simvastatin ≤20 mg, rosuvastatin ≤5 mg, pravastatin ≤40 mg, or fluvastatin ≤40 mg. Patients were excluded if they had a history of: 1) significant cardiovascular event or intervention within 12 weeks of screening; 2) type 1 diabetes; 3) significant heart failure (e.g., New York Heart Association functional class III or IV); 4) active or chronic hepatobiliary disease; or 5) systemic inflammatory condition or infection.
After initial screening, patients underwent baseline PET/CT imaging. Those with at least 1 vessel (either the aorta, right carotid, or left carotid) that had an average maximum target-to-background ratio (TBR) ≥1.6 were eligible for randomization (24). Prior statin therapy was discontinued. Patients were randomized 1:1 to either 1 × 10 mg Lipitor tablet (Pfizer, New York, New York) plus 1 × 80 mg atorvastatin matching placebo daily or 1 × 80 mg Lipitor atorvastatin tablet plus 1 × 10 mg atorvastatin matching placebo daily for 12 weeks. PET/CT images were obtained again 4 and 12 weeks following initiation of study drug.
FDG-PET/CT and Contrast-Enhanced CT Imaging
FDG-PET/CT imaging of the carotid arteries and ascending thoracic aorta was done using reproducible, validated methods (4,19). FDG was administered intravenously (10 mCi) after an overnight fast and imaging was performed 2 h later using PET/CT. An attenuation correction scan was obtained using voltage of 140 kVp. Thereafter, PET imaging was performed covering the neck and chest, with 15 min per bed position. Reconstruction of attenuation-corrected images was done using OSEM algorithm. All patients had a blood sugar concentration of <200 mg/dl at the time of imaging. Patients were provided specific instructions to follow a low-carbohydrate diet for 24 h before the test in order to suppress myocardial FDG uptake. Contrast-enhanced CT imaging was performed once (at baseline or at Week 4) to provide additional anatomical information to ensure that the same locations within the arterial segments are measured across time. CT parameters included tube voltage of 120 kVp, tube current of ∼750 mAs, section thickness of 0.75 mm, and pitch of 0.2 to 0.4, following injection of intravenous contrast. The average radiation exposure associated with each PET/CT was approximately 6 to 7 mCi, while the approximate radiation exposure associated with the CT angiogram was approximately 15 mCi.
Images were blinded to patient identifiers and imaging sequence and were analyzed at a central core lab (Massachusetts General Hospital) using Leonardo TrueD software (Siemens, Forchheim, Germany). The initial step of image analysis entailed delineating the pre-defined sections of the target vessels on the attenuation correction CT images (AC-CT) that are provided as part of the PET/CT dataset. To do so, the temporally blinded image sets (baseline, 4-, and 12-week PET/CT data) were inspected and the locations of interest were delineated by comparing the AC-CT images to ensure that the same locations are measured across time. For analysis of the carotid artery, where identification of the carotid bifurcation can be challenging using the AC-CT data alone, the contrast-enhanced CT images were used to facilitate the identification of the carotid bifurcation. To do so, the contrast-enhanced CT and AC-CT images were visually compared and the location of the carotid bifurcation was marked on the AC-CT. Thereafter, arterial FDG uptake was measured within pre-defined sections of the target vessels, guided by the AC-CT images (which are provided as automatically coregistered datasets accompanying the PET scans). For the aorta, measurements were made every 3 mm, starting 1 cm above the aortic valve annulus, continuing to the bottom of the aortic arch. For the right and left carotids, measurements were made every 3 mm, starting 2 cm below the carotid artery bifurcation, and continuing superiorly to 2 cm into the internal carotid artery. Measurements were made in the axial plane by drawing a region of interest around the artery wall (Fig. 1). The maximum standardized uptake value (SUV) was recorded for each region of interest. The SUV was calculated as a time- and dose-corrected tissue radioactivity divided by body weight. The TBR was calculated from the ratio of the SUV of the artery compared to background venous activity, derived from the superior vena cava (for correction of aortic values), or the internal jugular veins (for correction of carotid values). The intraobserver and interobserver variability of this approach (intraclass correlation coefficient [95% confidence interval]) are 0.92 (0.75 to 0.98), and 0.98 (0.94 to 1.00), respectively (25).
Arterial FDG uptake (TBR) was evaluated in the 3 target arterial locations (right and left carotid and aorta). The artery with the highest FDG uptake at baseline was identified as the index vessel, as previously described (20,24). Thereafter, index vessel FDG uptake (TBR) was calculated using 2 different approaches. The first approach (used to evaluate the primary endpoint) was to define the average of the maximum TBR activity within the most diseased segment (MDS) of the index vessel (MDS TBR). The MDS, in turn was defined as the 1.5-cm arterial segment, centered on the slice of artery demonstrating the highest FDG uptake at baseline, and calculated as a mean of maximum TBR values derived from 3 contiguous axial segments. The second approach was defined as the average of the maximum TBR activity for all of the axial segments that compose the index vessel (whole vessel TBR with highest mean of maximum TBR values at baseline).
The predefined primary endpoint was the relative mean change from baseline at 12 weeks in the arterial 18FDG-PET signal, assessed as a target to background ratio within the most diseased segment (MDS) of the index vessel. The pre-defined secondary endpoint evaluated the relative change in the whole vessel TBR within the index vessel. Exploratory endpoints included: 1) assessment of treatment effect of relative change in MDS and whole vessel TBR at 4 weeks; 2) assessment of treatment effects on carotid arteries and aorta separately, at 4 and 12 weeks; and 3) assessment of the relationship between lipid parameters and TBR.
The primary objective of the study was to determine whether 12 weeks of treatment with atorvastatin 80 mg results in a greater reduction from baseline in the FDG-PET TBR of the MDS when compared to 10 mg of atorvastatin. Relative mean changes from baseline in FDG-PET MDS TBR at 12 weeks were compared between treatments using an analysis of covariance model including terms for treatment and baseline covariate (predefined in the initial protocol), and prior statin use (predefined prior to unblinding). The Week 12 geometric mean fold change from baseline in the mean of the maximum TBR of the MDS was estimated for each dose and assessed for statistical significance between groups. Within group relative changes from baseline were also assessed as supportive data; hence no multiplicity adjustment was performed. Data were log transformed as pre-specified in the statistical analysis plan. Differences between treatments were assessed based on the observed data values, without adjusting for missing data. Additional sensitivity analyses were performed using multiple imputation to attempt to correct for any missing data (Online Appendix). Week 4 data were similarly analyzed. Associations were tested using Pearson's correlation coefficient. Multiplicity adjustments were not applied for secondary comparisons of interest; as such, nominal p values are reported for all comparisons. Post hoc treatment group comparisons of baseline variables were performed using either a Wilcoxon rank sum test for continuous endpoints or Fisher's exact test for binary endpoints.
The sample size of 32 patients per treatment group had 77% power to yield a statistically significant difference (alpha = 0.05, 1-sided) in MDS TBR between treatments if the true underlying difference between treatments is 0.13 (∼7% of expected baseline mean TBR 1.9).
A total of 163 patients were screened and 83 patients were randomized (Fig. 2). Exclusion was due to TBR <1.6 (n = 16), HbA1c >8% (n = 11), fasting plasma glucose >200 mg/dl (n = 7), and excluded medications (n = 5). Twelve patients discontinued early due to adverse experiences (n = 6), withdrawal of consent (n = 2), protocol deviation, or other reason (n = 4). Patient demographics were generally comparable across treatment groups as shown in Table 1.
Effect of atorvastatin on atherosclerotic plaque activity
Primary endpoint analysis, evaluating change from baseline in the TBR within the MDS of the index vessel after 12 weeks demonstrated a statistically significant relative reduction from baseline in the atorvastatin 80 mg group versus atorvastatin 10 mg groups (Table 2, Fig. 3). At 12 weeks a greater reduction in FDG uptake was observed in patients randomized to atorvastatin 80 versus 10 mg, with significant additional reduction in FDG uptake within the MDS (10.6% [2.2% to 18.3%]; p = 0.01) Arterial FDG uptake measurements made at 4 weeks were generally consistent with those seen at 12 weeks (Fig. 3). Sensitivity analyses using multiple imputation showed generally consistent results, with a reduction from baseline in the atorvastatin 80 mg group of 8.8% (p < 0.05) (Online Table 1). The impact of baseline measures on the amount of MDS reduction was modest (estimated correlation = –0.40, p = 0.001); and the impact was more pronounced for the 80 mg group than the 10 mg group. Consistency of the effect of statins on arterial FDG uptake was observed when the carotid arteries (as average of right and left carotid) were examined separately from the aorta (Online Table 2).
Relationship between lipids and vascular inflammatory activity
Significant, dose-dependent reductions from baseline were observed for total plasma cholesterol, LDL, and triglycerides (Fig. 4A, Online Table 3). No statistically significant, dose-dependent increases in high-density lipoprotein cholesterol were observed. While a trend between baseline LDL cholesterol concentrations and baseline MDS TBR was observed (r = 0.21, p < 0.1) (Fig. 4B), there was no significant correlation between the changes in LDL cholesterol and changes in TBR during the study (r = 0.04, p = 0.74) (Fig. 4C). Furthermore, there were no significant relationships between C-reactive protein (CRP) and MDS TBR, either at baseline (r = 0.11, p = 0.41), or for change from baseline (r = 0.05, p = 0.73). Likewise, there were no significant relationships between CRP and LDL cholesterol, either at baseline (r = 0.15, p = 0.23) or for change from baseline (r = 0.11, p = 0.40).
Effect of statin intensification
A post hoc analysis of the subset of patients in the 80 mg atorvastatin arm who were receiving a low-dose statin prior to randomization (n = 18) showed a significant reduction in MDS TBR at 12 weeks versus baseline (13.4% reduction [5.6% to 20.6%]; p < 0.01).
Several novel findings were observed in this first multicenter feasibility study evaluating the effects of statins on vascular inflammation determined by FDG-PET. First, we observed a dose response in the reduction in FDG uptake between the high- and low-dose statin groups. Second, much of the observed reduction in plaque activity was apparent as early as 4 weeks after randomization to atorvastatin 10 versus 80 mg. Taken together, these findings provide evidence of a rapid reduction in vascular inflammation with statin therapy and provide new insights regarding the graded reductions in vascular plaque activity as it relates to increases in statin doses. The results also confirm the ability of PET imaging as a tool to detect changes in vascular inflammation early in the course of treatment, something not as well validated with other noninvasive imaging methods (26).
Comparison of study findings with clinical endpoint trial experience with statins
Atorvastatin therapy significantly reduces major cardiovascular events, at both low (10 mg) and high (80 mg) doses (27–29). In addition, an incremental benefit for 80 mg over 10 mg atorvastatin on the rate of cardiovascular events was observed in the TNT (Treating to New Targets: Time to Occurrence of Major Cardiovascular Events Trial) trial (27,28). Similar to cardiovascular events, an incremental benefit of statin intensification was observed on plaque inflammation measured using FDG-PET imaging as reported in the current study. These data support the hypothesis that statin therapy may result in a reduction in cardiovascular benefit in part due to a rapid reduction in arterial inflammation.
Comparison with circulating biomarkers
In the current trial, the treatment effects of atorvastatin on circulating lipid parameters (LDL cholesterol in particular) were consistent with previous reports. It is interesting that whilst the changes in both FDG uptake and circulating LDL cholesterol were directionally concordant and followed a similar time course, there did not appear to be a significant correlation between the 2 parameters at the individual patient level. This is not unlike findings from prior larger studies that also showed modest correlations between inflammatory biomarkers and LDL. For example, in the PROVE-IT–TIMI 22 (Pravastatin or Atorvastatin Evaluation and Infection Therapy–Thrombolysis in Myocardial Infarction 22) trial, the relationship between achieved LDL cholesterol and achieved CRP levels was modest to nonexistent for the statins employed in that study (r = 0.04, p = 0.07 for pravastatin; r = 0.15, p = 0.001 for atorvastatin) (30). Accordingly, it is possible that statins' effect on vessel wall inflammation may be, at least in part, independent of changes in LDL cholesterol (“pleiotropic effects”). However, it is also possible that the current trial lacked the statistical power to provide evidence of an effect related to changes in LDL cholesterol.
Comparison with prior PET/CT studies
The magnitude of the change in FDG uptake after treatment with the low-dose atorvastatin in this study is similar to that seen in the open-label study using low-dose evaluating simvastatin versus diet (20), therefore, there is a consistent reduction in the PET signal associated with low-dose statins. On the other hand, dalcetrapib, a cholesteryl ester transfer protein modulator was not associated with changes in the co-primary FDG-PET endpoint, index vessel MDS as used in this study (p = 0.37), though a near-significant reduction in the exploratory measure of carotid MDS TBR, was also noted (p = 0.07) (24). Dalcetrapib was subsequently found not to be effective in reducing cardiovascular events. Future studies should evaluate, using drugs from other therapeutic classes, whether there is a consistent relationship between changes in the PET signal and clinical benefit.
It remains unclear whether lowering plaque inflammation is a mechanism by which statin therapy decreases cardiovascular events or even whether reducing arterial inflammation translates into clinical benefit at all. Future studies wherein imaging is embedded within clinical endpoint trials will be needed to evaluate the relationship between plaque inflammation and clinical events. Should a relationship between reductions in plaque inflammation and reductions in clinical cardiovascular endpoints be proven, then targeting plaque inflammation may find a central role in cardiovascular therapy (31).
In this study, larger treatment effects were observed within the MDS compared to the whole vessel. This was not unexpected, since the whole vessel is an admixture of diseased and relatively less diseased sections. Accordingly, the response to effective therapy is expected to be heterogeneous within the whole vessel. On the other hand, the MDS represents the region of greatest inflammation, hence theoretically represents a location that may be more susceptible to treatment effect. Nonetheless, in the current study, the other endpoints were all generally consistent and directionally concordant with the changes seen in the MDS, attesting to the robustness of the results themselves. It is likely that for future studies, the MDS should serve as a key endpoint to compare randomized treatment groups.
First, though the study employed use of an active comparator arm, no true placebo control group is included, due to ethical considerations. Second, while the current trial showed significant effects of statins on the vessel wall at 4 and 12 weeks, it was not designed to evaluate if these effects persist over longer durations of treatment. Further work in longer term trials will be required to address the question of durability of treatment effects. Third, a substantial number of subjects were unable to provide complete data sets for analysis (consistent with prior imaging trial experiences). Nonetheless, analyses of the data that included and excluded multiple imputation methods demonstrated similar results, thus enhancing confidence in the study results.
This first multicenter trial evaluating the effects of statins on vascular inflammation imaged by FDG-PET has provided several important insights, including the observation of a significant dose response in the reduction in FDG uptake between the high- and low-dose statin groups along with an observation that reductions in plaque activity are apparent as early as 4 weeks after randomization to statins. These observations are consistent with the findings from large-scale event-driven trials of low-versus high-dose statins. Importantly, the trial data lend strong support to the potential use of FDG-PET imaging as a tool for early assessment of treatment effect in patients already taking statins.
The authors acknowledge the following investigators and sites for clinical and imaging assistance: Jacqueline Brunetti, MD, Nathaniel Lebowitz, MD, Holy Name Medical Center, Department of Radiology, Teaneck, New Jersey; Neil J. Fraser, MD, FACP, Troy Internal Medicine Research, Troy, Michigan; Peter A. McCullough, MD, Ching Yee Wong, MD, William Beaumont Hospital, Beaumont Health Center, Royal Oak, Michigan; Brooke J Nevins, MD, Bronxville, NY; Eli M. Roth, MD, FACC, CPI, Sterling Research Group, Ltd., Cincinnati, Ohio; James M. Rhyne, MD, Lipid Center, Statesville, North Carolina; Joel Rubenstein, MD, Newton-Wellesley Cardiology, Newton, Massachusetts; Cezar S. Staniloae, MD, Gotham Cariovascular Research, PC, New York, New York. Editorial assistance was provided by Wendy Horn, PhD, of Insight Communications Group, Westfield, New Jersey. This assistance was funded by Merck Sharp & Dohme Corporation, a subsidiary of Merck & Co., Inc., Whitehouse Station, New Jersey.
Drs. Tawakol and Fayad received consulting fees, and their institutions received grants from Roche and Merck Sharp & Dohme Corporation, a subsidiary of Merck & Co. Dr. Farkouh received support for travel and his institution received grants from Merck Sharp & Dohme Corporation. Drs. Mogg, Alon, Klimas, Dansky, Nunes, Beals, and Shankar are employees of Merck Sharp & Dohme Corporation, a subsidiary of Merck & Co. Drs. Mogg, Alon, Dansky, Nunes, and Beals own stock in Merck Sharp & Dohme Corporation. Dr. Rudd is supported by the NIHR Cambridge Biomedical Research Center. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Tawakol and Fayad contributed equally to this work.
- Abbreviations and Acronyms
- attenuation correction
- C-reactive protein
- low-density lipoprotein
- most diseased segment
- positron emission tomography-computed tomographic imaging
- standardized uptake value
- target-to-background ratio
- Received March 4, 2013.
- Revision received March 27, 2013.
- Accepted April 9, 2013.
- American College of Cardiology Foundation
- Rudd J.H.,
- Warburton E.A.,
- Fryer T.D.,
- et al.
- Tawakol A.,
- Migrino R.Q.,
- Bashian G.G.,
- et al.
- Rodriguez-Prados J.C.,
- Traves P.G.,
- Cuenca J.,
- et al.
- Kubota R.,
- Kubota K.,
- Yamada S.,
- Tada M.,
- Ido T.,
- Tamahashi N.
- Rudd J.H.,
- Myers K.S.,
- Bansilal S.,
- et al.
- Kim T.N.,
- Kim S.,
- Yang S.J.,
- et al.
- Rogers I.S.,
- Nasir K.,
- Figueroa A.L.,
- et al.
- Rominger A.,
- Saam T.,
- Wolpers S.,
- et al.
- Rudd J.H.,
- Myers K.S.,
- Bansilal S.,
- et al.
- Tahara N.,
- Kai H.,
- Ishibashi M.,
- et al.
- Ray K.K.,
- Cannon C.P.,
- McCabe C.H.,
- et al.,
- for the PROVE IT-TIMI 22 Investigators
- Rudd J.H.F.,
- Myers K.S.,
- Bansilal S.,
- et al.
- Joshi F.,
- Lindsay A.C.,
- Obaid D.R.,
- Falk E.,
- Rudd J.H.
- Westerterp M.,
- Koetsveld J.,
- Yu S.,
- et al.