Nicholls
M. John Chapman, Raimund Erbel, Joel S. Raichlen, Kiyoko Uno, Yu Kataoka and Stephen J.
Rishi Puri, Steven E. Nissen, Peter Libby, Mingyuan Shao, Christie M. Ballantyne, Phillip J. Barter,
Statin Therapy
Coronary Atheroma Regression and Cardiovascular Events Following Maximally Intensive
C-Reactive Protein, but not Low-Density Lipoprotein Cholesterol Levels, Associate with
Print ISSN: 0009-7322. Online ISSN: 1524-4539
Copyright © 2013 American Heart Association, Inc. All rights reserved.
is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation
published online September 16, 2013;Circulation.
http://circ.ahajournals.org/content/early/2013/09/16/CIRCULATIONAHA.113.004243
World Wide Web at:
The online version of this article, along with updated information and services, is located on the
http://circ.ahajournals.org//subscriptions/
is online at: Circulation Information about subscribing to Subscriptions:
http://www.lww.com/reprints
Information about reprints can be found online at: Reprints:
document. Permissions and Rights Question and Answer available in the
Permissions in the middle column of the Web page under Services. Further information about this process is
Once the online version of the published article for which permission is being requested is located, click Request
can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.Circulation
Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:
at University of Michigan--Ann Arbor on September 18, 2013http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.113.004243
1
C-Reactive Protein, but not Low-Density Lipoprotein Cholesterol Levels,
Associate with Coronary Atheroma Regression and Cardiovascular Events
Following Maximally Intensive Statin Therapy
Running title: Puri et al.; Statins, CRP, plaque progression and MACE
Rishi Puri, MBBS1; Steven E. Nissen, MD1; Peter Libby, MD2; Mingyuan Shao, MS3;
Christie M. Ballantyne, MD4; Phillip J. Barter, MBBS, PhD5; M. John Chapman, PhD, DSc6;
Raimund Erbel, MD7; Joel S. Raichlen, MD8; Kiyoko Uno, MD, PhD1; Yu Kataoka, MD1;
Stephen J. Nicholls, MBBS, PhD1,9
1Dept of Cardiovascular Medicine, Cleveland Clinic, Cleveland, OH; 2Cardiovascular Division,
Brigham and Women’s Hospital, Boston, MA; 3C5Research, Cleveland Clinic, Cleveland, OH;
4Section of Cardiovascular Research, Baylor College of Medicine, and the Methodist DeBakey
Heart and Vascular Center, Houston, TX; 5Centre for Vascular Research, University of New
South Wales, Sydney, Australia; 6INSERM Dyslipidaemia and Atherosclerosis Research Unit,
Hôpital de la Pitié, Paris, France; 7West German Heart Center, Essen, Germany; 8AstraZeneca,
Wilmington, DE; 9South Australian Health and Medical Research Institute, University of
Adelaide, Adelaide, Australia
Address for Correspondence:
Stephen J. Nicholls, MB, BS, PhD
South Australian Health and Medical Research Institute
Level 9, 121 King William St
Adelaide, SA, 5001, Australia
Tel: 61-8-81164432
Fax: 61-8-81164432
E-mail: stephen.nicholls@sahmri.com
Journal Subject Codes: Atherosclerosis:[135] Risk factors, Atherosclerosis:[150] Imaging,
Atherosclerosis:[90] Lipid and lipoprotein metabolism, Treatment:[122] Secondary prevention
Raimund Erbel, MD ; Joel S. Raichlen, MD ; Kiyoko Uno, MD, PhD ; Yu Katataoaoaokaka, , MDMD ;;
Stephen J. Nicholls, MBBS, PhD1,9
1DeDeptptp oooff f CaCaardrdr iooovavavascs ular Medicine, Cleveland CCliliininn c, Cleveland, OHH;; 2CaCC rdiovascular Division,
BrBrBrigigighhaham m anananddd WoWoW mmen’s Hosppital, Boston, MA; 333C5C5Research, Clleveve ellananddd Clinic, Cleveland, OH;
444Seeectc ion off CCCararardididiovovvasasascucuculalalarr r ReReResesesearararchchh,, BaBaB ylylylororor CCoollleeegee ooof f f MeMeMedidiccicinenee, , anananddd tthe e MeMeMethththododo isisst tt DeDeDeBaBaBakekek y
HHeHeart and VaVassscuuularr CCCenteter,r, HHououustststonnn, TTX;;; 5CCennntrree fororor VVVaaasccculalaar RReseeearrcrch,h,h, UUUniniveveersssitityyy oofof NNewww
SoSoSoutuu h Walelees,ss SSydyddneyy, Ausustrtralala iaia; 66ININI SESEERMMM DDysysliippidadadaeememiiaa aanddd AAAtherrrosssclcleere oososisi RRReeseseaaarccch UnUnUnit,
HôHôôpipipitatatalll dedede lla a PiPiPititié,é,é PPPararisisis,, Frrananancee;;; 777WeWeestst GGGererermamaman nn HeHeHeararart t CeCeCentntntereer, EsEsEssessen,n,n, GGGerere mamamanynyny; ; ; 888AsAsAsttrtraZaZaZenenenececca,a,a
WiWiWilmlmlmininingtgtgtononon,, DEDEDE;;; 9SoSoSoutututhhh AuAuAustststrararalililiananan HHHeaeaealtltlthhh anananddd MeMeMedididicacacalll ReReResesesearararchchch IIInsnsnstitititutututetete,, UnUnUnivivivererersisisitytyty ooofff
AdAdAdelllaiaiaidedede,, AdAdAdelelelaiaiaidedede,, AuAuAustststrararalililiaaa
at University of Michigan--Ann Arbor on September 18, 2013http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.113.004243
2
Abstract
Background—Baseline C-reactive protein (CRP) levels predict major adverse cardiovascular
events [MACE; death, myocardial infarction, stroke, coronary revascularization and
hospitalization for unstable angina]. The association between changes in CRP levels with plaque
progression and MACE in the setting of maximally intensive statin therapy is unknown.
Methods and Results—SATURN employed serial intravascular ultrasound (IVUS) measures of
coronary atheroma volume in patients treated with rosuvastatin 40 mg or atorvastatin 80 mg for
24 months. The treatment groups did not differ significantly in the change from baseline of
percent atheroma volume (PAV) on IVUS, CRP-modulating effects or MACE rates, thus
allowing for a (prespecified) post-hoc analysis to test associations between changes in CRP
levels with coronary disease progression and MACE. Patients with non-increasing CRP levels
(n=621) had higher baseline [2.3 (1.1, 4.7) vs. 1.1 (0.5, 1.8) mg/L, p<0.001] and lower follow-up
CRP levels [0.8 (0.5, 1.7) vs. 1.6 (0.7, 4.1) mg/L, p<0.001] versus those with increasing CRP
levels (n=364). Multivariable analysis revealed a non-increasing CRP level to independently
associate with greater PAV regression (p=0.01). While the (log) change in CRP did not associate
with MACE [HR 1.18 (95% CI 0.93, 1.50), p=0.17], the (log) on-treatment CRP associated
significantly with MACE [HR 1.28 (95% CI 1.04, 1.56, p=0.02). On-treatment LDL-C levels did
not correlate with MACE [HR 1.09 (95% CI 0.88, 1.35, p=0.45).
Conclusions—Following 24 months of potent statin therapy, on-treatment CRP levels associated
with MACE. Inflammation may be an important driver of residual cardiovascular risk in patients
with coronary artery disease despite aggressive statin therapy.
Clinical Trial Registration Information—ClinicalTrials.gov. Identifier: NCT000620542.
Key words: statin, inflammation, low-density lipoprotein cholesterol, intravascular ultrasound,
atherosclerosis, C-reactive protein
y p g gg
n=621) had higher baseline [2.3 (1.1, 4.7) vs. 1.1 (0.5, 1.8) mg/L, p<0.001] and d lolloweww rr r fofofollllllowowow-u- pp
CRP levels [0.8 (0.5, 1.7) vs. 1.6 (0.7, 4.1) mg/L, p<0.001] versus those with increasing CRP
evels (n=364). Multivariable analysis revealed a non-increasing CRP level to independently
asasssosocciciatate wiwiwiththth ggrereatatere PPAVAV regegreessioi n n (pp=0.01)1).. WhWhW ile ththe e (l( ogg)) chchc anngege iin CRRPP did non t t asa sociate
wwithhh MACE [HHR RR 111.181818 ((959595%% % CICIC 00.9.9.93,3,3 11.5550), p===0.1177]], thhheee (l(logogog)) ononn-tttreeatatmemem ntnt CCCRPRP aaasssssococciaaatetteddd
iigngng ififi iccantly y y wiwiw thhh MMAACACEE [HHHR R R 1.1.28288 ((9955%% % CCII 111.0.044, 11.5.566,6, pp==0..002)2). OOnOn-trrreaatatmmmentntn LLDLDLDL-CCC llevevelells did
not t cocorrr elelatatee wiwitht MMACA EE [H[HR R 11.0909 (95% % CICI 00.888,8 1.335,5 pp=0=0.4.45)..
Conclusionnsss——FoFoFollllllowowowininng g 242424 mmononnththhss ofoff ppotototenenentt t stststatatininin tttheerarrapypyy, onoon-trereeatatatmememennnt tt CRCRCRPP P leleevevev lslsl associateddd
at University of Michigan--Ann Arbor on September 18, 2013http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.113.004243
3
Statin-mediated cholesterol lowering improves clinical outcomes. But despite this, many
individuals demonstrate atheroma progression,1 and many such patients experience further
cardiovascular events. Randomized clinical trials employing the best contemporary medical
therapies demonstrate, at most, relative risk reductions of the order of 30-40%. This highlights a
significant degree of residual cardiovascular risk experienced by a large proportion of individuals
prescribed statins for secondary clinical prevention.2 Accordingly, there is a need to identify
novel targets and therapeutic strategies to effectively combat this residual risk.
Multiple avenues of research implicate inflammation as a factor involved in the initiation,
progression and instability of atherosclerotic plaques.3, 4 Elevations of serum inflammatory
biomarkers consistently associate with the risk of experiencing a cardiovascular event.5, 6 C-
reactive protein (CRP), in particular, has been extensively investigated, and the prognostic utility
of baseline CRP levels is now considered by some, to be equivalent to that of serum cholesterol
levels.7, 8 While many benefits of statins are attributed to cholesterol lowering, both coronary
atheroma regression9 and clinical event reductions10 have also been shown independently to
correlate with statin-mediated CRP lowering. Yet there remains an ongoing debate as to whether
the effects of statin-mediated CRP lowering are simply a function of parallel changes in low-
density lipoprotein cholesterol (LDL-C), or follow a separate mechanistic pathway. The
prognostic benefit of CRP in statin-treated patients who achieve LDL-C levels <70 mg/dL, is
uncertain.11
SATURN (Study of Coronary Atheroma by Intravascular Ultrasound: Effect of
Rosuvastatin Versus Atorvastatin; ClinicalTrials.gov number, NCT000620542) was the first
head-to-head comparison of the anti-atherosclerotic effect of two potent statin agents, each
prescribed at their maximal approved dose, on coronary atherosclerosis progression.12 Treatment
biomarkers consistently associate with the risk of experiencing a cardiovascular eeevenee tt..5,5, 5 666 CCC--
eactive protein (CRP), in particular, has been extensively investigated, and the prognostic utility
off bbbasasaselelelininineee CRCRC PP lelelevvels is now considered by sommeme,, to be equivaleeentnn to o thththata of serum cholesterol
eeveeelsl .7, 8 Whiilele mmmananyy bebeenenenefifiitststs oof f ststs aaatitinsns aree attttribbubutted totoo cchohoh lelessttterrrol l lololowweeriringngng,, bobooththth ccoorronononararyy y
attheheheroroomamam rrregegegreresssssiooonn99 aanndd ccliinnin cacaalll evevevenennt t rerereduduductttioioionnsns1100 hhhavavveee alaalsooo bbeeeee nnn shshs owowownn ininindededepepependndndennntlyy y totoo
correlate withthh sstatat tititin-n-n mememedid atatatededed CCCRPRPR lllowowwerererinining.g.g. YYYetetet thththerere e e rerer mamamainnns s s ananan ooongngngoioioingngng ddebebebatatate e e asasas to whetherrr
at University of Michigan--Ann Arbor on September 18, 2013http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.113.004243
4
with both agents was associated with regression of coronary atheroma, although there was no
significant difference between the two groups in the primary efficacy end point of change in
percent atheroma volume (PAV); nor was there a difference in safety or clinical event rates
between each treatment group. The primary objective of this prespecified post hoc analysis was
to examine the prognostic effect of changes in CRP levels on both coronary atheroma
progression and major adverse cardiovascular events (MACE) in patients following 24 months of
maximally intensive statin therapy.
Methods
Patient selection
The design of SATURN has been previously described.13 Briefly, patients with angiographically
demonstrable coronary disease and 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. Subjects underwent intravascular
ultrasound (IVUS) imaging of a coronary artery at baseline and after 104 weeks of treatment.
Acquisition and analysis of intravascular coronary imaging
The presence of at least a single lumen stenosis of >20% angiographic diameter stenosis severity
in an epicardial coronary artery at the time of a clinically indicated coronary angiogram was
necessary for enrolment eligibility. IVUS was performed at baseline in a single, native coronary
artery with no lumen stenosis of >50% severity, which had not undergone revascularization and
was not considered to be the culprit vessel of a prior myocardial infarction (MI). Images were
screened by the Atherosclerosis Imaging Core Laboratory of the Cleveland Clinic Center for
Clinical Research for quality, and those patients whose baseline imaging met these requirements,
Patient selection
The design of SATURN has been previously described.13 Briefly, patients with angiographically
deemomomonsnsnsttrtrababablelele cororroononary disease and LDL-C <11666 mmg/dL followinng gg a 2-2--wweweek treatment period
wwithhh atorvastaatitiinnn 4440 mmmg ororr rrosososuuvuvasasastataatitinnn 220 mgmgmg daiiilyyy wwerereree reree-rrrananddodomimizezeedd anannddd ttrtreaeattet d d d fofofor 242424--
momoontntnthshshs wwititth h h atatororrvavaaststaatatinnn 880 mgmgmg ooor rr rrorosusuuvavavastststaaatininn 44000 mgmgmg dddaaan ilililyyy. SSSubububjejeectctctss uunundederwrwrwenene t t ininintttravavvasasscuculllarr
ultrasound ((IVIVVUSUSUS)) ) imimimagagaginnggg ofofo aaa cccororrooonaaaryryry aartrtr ererery y y atatat bbbasasa elellininine ee ananand d d afaa teteer r r 101010444 wewew ekekksss ofofof ttrerer atment. ff
at University of Michigan--Ann Arbor on September 18, 2013http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.113.004243
5
were eligible for randomization. Following 104 weeks of treatment, patients underwent a second
IVUS of the same artery. Anatomically matched arterial segments were selected for analysis on
the basis of proximal and distal landmarks. Cross-sectional images spaced 1mm apart were
selected for analysis, with lumen and external elastic membrane (EEM) leading edges defined by
manual planimetry. Plaque area was determined as the area between these leading edges. Percent
atheroma volume (PAV), a measure of plaque burden, was calculated as previously described.14
Change in plaque burden was calculated as the PAV at 104 weeks minus the corresponding PAV
at baseline. Plaque regression was defined as any decrease in PAV from baseline. The post-hoc
analyses presented here pooled results from both treatment groups; as in SATURN, they did not
differ in the primary efficacy endpoint of the change in PAV from baseline, and the average on-
treatment CRP levels were not different between the rosuvastatin- and atorvastatin-treated
groups.
Statistical analysis
Continuous variables were reported as mean ± SD if normally distributed and as median
(interquartile range) if non-normally distributed. Demographics, baseline clinical characteristics,
follow-up medications, baseline ultrasonic and laboratory biochemical data were compared
between those who had an increase in average on-treatment CRP levels from baseline, against
those whose on-treatment CRP levels did not increase. Two-sample t-tests were used for
normally distributed continuous variables, Wilcoxon rank-sum tests for non-normally distributed
continuous variables, and chi-square tests for categorical variables. Serial changes in IVUS
measurements were analyzed by analysis of variance (ANOVA), adjusting for their baseline
counterparts, and were reported as least-squares mean ± SE. Given the significant difference
observed for change in PAV between those whose CRP increased versus those whose CRP did
differ in the primary efficacy endpoint of the change in PAV from baseline, and d tththe ee avvverereragagageee onon-
reatment CRP levels were not different between the rosuvastatin- and atorvastatin-treated
grgrououupspsps..
SStStatattisi tical annalalyyysiiis
CoCoontntntinininuouoususs vvvarariaiaabblleses wweereree reeepopoportrtrtededed aas s mememeananan ±±± SSSDDD iifif nnnororrmamamally y y didiisttririribububuttteddd anannd d asasa mmmeeediaiaan n
interquartilee rrrananangegege))) ififf nnnonnn-n-nnoroo mamamallllyy y diiiststs ririribububuteteted.d.d DDDememmogoggrararaphphphicccs,s,s, bbasasaseleleliinenene ccclilil niniicacacal l chchcharara acteristics,,
at University of Michigan--Ann Arbor on September 18, 2013http://circ.ahajournals.org/Downloaded from
DOI: 10.1161/CIRCULATIONAHA.113.004243
6
not increase, a multivariable ANOVA analysis was undertaken to assess for its association with
changes in CRP levels after controlling for clinical factors. Partial correlation analyses were
performed to assess for relationships between change in CRP and lipid parameters, or change in
PAV. A multivariable logistic regression model was constructed for defining important
predictors of any reduction in CRP from baseline.
A multivariable Cox proportional hazards model was derived to identify clinical factors
that were associated with MACE, (defined as cardiovascular death, non-fatal MI, non-fatal
stroke, coronary revascularization or hospitalization for unstable angina), along with examining
the effect of CRP (change in CRP and on-treatment CRP levels) on MACE. To create these
multivariable models, demographic data and clinical characteristics were entered into a
multivariable logistic or Cox model (with CRP forced into this model) for variable selection with
bootstrap re-sampling (1000 iterations and a P value criterion of 0.1 for retention). Those
variables having a 40% or more probability of retention were entered into a second logistic or
Cox model with the stepwise model selection procedure. The significance level to enter and keep
a variable was set at 0.05. For the Cox model, the risk factors met the proportional hazards
assumption by plotting the log of the negative log of their estimated survival distribution, log(–
log(S(t))), versus time, and the linearity assumption was assessed and satisfied for all continuous
variables by plotting the log hazard ratio against the continuous variable as well as testing the
linear hypotheses. The selected covariates formed the covariate set for the final multivariable
logistic model for on-treatment CRP or the final multivariable Cox model for MACE. Given the
significant association between on-treatment CRP levels and MACE, further evaluations were
conducted by performing log-rank tests with Kaplan-Meier figures and unadjusted and adjusted
Cox proportional hazards regressions with hazard ratio (HR) and associated 95% confidence
multivariable models, demographic data and clinical characteristics were entered d d innntotot aaa
multivariable logistic or Cox model (with CRP forced into this model) for variable selection withd
boootototstststrararapp rerere-s-ssammmplplplining (1000 iterations and a P vavavaluluee criterion of 000.1 ffororor rrrete ention). Those
vvariiiaba les haviingng aa 40%0%0% ooorr r momomorrere ppprororobababibillity offf retteennntiooon n weweereee eennteerreded iiintntoo a a a seseecocondndnd lllogoggisssti