Chester Clinic

 

Premature atherosclerosis in pediatric systemic lupus erythematosus: risk factors for increased carotid intima-media thickness in the atherosclerosis prevention in pediatric lupus erythematosus cohort

ARTHRITIS & RHEUMATISM Vol. 60, No. 5, May 2009, pp 1496–1507 DOI 10.1002/art.24469 2009, American College of Rheumatology Premature Atherosclerosis in Pediatric Systemic Lupus Erythematosus Risk Factors for Increased Carotid Intima-Media Thickness in the Atherosclerosis Prevention in Pediatric Lupus Erythematosus Cohort Laura E. Schanberg,1 Christy Sandborg,2 Huiman X. Barnhart,1 Stacy P. Ardoin,1 Eric Yow,1 Gregory W. Evans,3 Kelly L. Mieszkalski,1 Norman T. Ilowite,4 Anne Eberhard,5 Deborah M. Levy,6 Yukiko Kimura,7 Emily von Scheven,8 Earl Silverman,9 Suzanne L. Bowyer,10 Lynn Punaro,11 Nora G. Singer,12 David D. Sherry,13 Deborah McCurdy,14 Marissa Klein-Gitelman,15 Carol Wallace,16 Richard Silver,17 Linda Wagner-Weiner,18 Gloria C. Higgins,19 Hermine I. Brunner,20 Lawrence Jung,21 Jennifer B. Soep,22 and Ann Reed,23 for the Atherosclerosis Prevention in Pediatric Lupus Objective. To evaluate risk factors for subclinical
ments of carotid intima-media thickness (CIMT) as
atherosclerosis in a population of patients with pediat-
part of the Atherosclerosis Prevention in Pediatric
ric systemic lupus erythematosus (SLE).
Lupus Erythematosus (APPLE) trial. SLE disease mea-
Methods. In a prospective multicenter study, a
sures, medications, and traditional risk factors for
cohort of 221 patients underwent baseline measure-
atherosclerosis were assessed. A standardized protocol
was used to assess the thickness of the bilateral common
ClinicalTrials.gov identifier: NCT00065806.
carotid arteries and the mean maximal IMT of 12
Supported by the NIH (contract N01-AR-2-2265 from the segments. Univariable analysis identified potential as-
National Institute of Arthritis and Musculoskeletal and Skin Diseases)and in part by the Duke Edna and Fred L. Mandel Jr. Center for sociations with CIMT, which were examined in multiva-
Hypertension and Atherosclerosis. Dr. Singer's work was supported in riable linear regression modeling.
part by the NIH (grant UL1-RR-024989 from the National Center forResearch Resources). Pfizer provided the study drug for the Athero-sclerosis Prevention in Pediatric Lupus Erythematosus (APPLE) trial.
western University Feinberg School of Medicine, Chicago, Illinois; 1Laura E. Schanberg, MD, Huiman X. Barnhart, PhD, Stacy 16Carol Wallace, MD: Seattle Children's Hospital, Seattle, Washing- P. Ardoin, MD, Eric Yow, MS, Kelly L. Mieszkalski, MA, CCRC: ton; 17Richard Silver, MD: Medical University of South Carolina, Duke University Medical Center, Durham, North Carolina; 2Christy Charleston; 18Linda Wagner-Weiner, MD: University of Chicago, Sandborg, MD: Stanford University School of Medicine, Palo Alto, Chicago, Illinois; 19Gloria C. Higgins, MD, PhD: Nationwide Chil- California; 3Gregory W. Evans, MA: Wake Forest University School dren's Hospital, Columbus, Ohio; 20Hermine I. Brunner, MD, MS: of Medicine, Winston-Salem, North Carolina; 4Norman T. Ilowite, Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio; MD: Albert Einstein College of Medicine, Bronx, New York; 5Anne 21Lawrence Jung, MD: Creighton University Medical Center, Omaha, Eberhard, MD: Schneider Children's Hospital, New Hyde Park, New Nebraska; 22Jennifer B. Soep, MD: The Children's Hospital, Denver, York; 6Deborah M. Levy, MD, MS: Columbia University Medical Colorado; 23Ann Reed, MD: Mayo Clinic, Rochester, Minnesota.
Center, New York, New York; 7Yukiko Kimura, MD: Hackensack Drs. Schanberg and Sandborg contributed equally to this University Medical Center, Hackensack, New Jersey; 8Emily von Scheven, MD, MS: University of California at San Francisco; 9Earl Dr. Schanberg has received consulting fees from Pfizer (less Silverman, MD: Toronto Hospital for Sick Children, Toronto, On- than $10,000). Mr. Evans has receiving consulting fees from AstraZen- tario, Canada; 10Suzanne L. Bowyer, MD: Indiana University School eca, Sanofi-Aventis, Schering-Plough, and Merck (less than $10,000 of Medicine, Indianapolis; 11Lynn Punaro, MD: Texas Scottish Rite Hospital for Children, Dallas; 12Nora G. Singer, MD: University Address correspondence and reprint requests to Laura E.
Hospitals/Case Medical Center, Cleveland, Ohio; 13David D. Sherry, Schanberg, MD, Duke University Medical Center, Box 3212, Durham, MD: Children's Hospital of Philadelphia, Philadelphia, Pennsylvania; NC 27710. E-mail: [email protected].
14Deborah McCurdy, MD: University of California Los Angeles Submitted for publication September 5, 2008; accepted in School of Medicine; 15Marissa Klein-Gitelman, MD, MPH: North- revised form January 31, 2009.
PREMATURE ATHEROSCLEROSIS IN PEDIATRIC SLE Results. Based on the mean-mean common or the
long-term cardiovascular complications in pediatric- mean-max CIMT as the dependent variable, univariable
onset SLE presents a particularly attractive target for analysis showed significant associations of the following
intervention, with the possibility of significantly improv- variables with increased CIMT: increasing age, longer
ing the quality of life and increasing survival over many SLE duration, minority status, higher body mass index
(BMI), male sex, increased creatinine clearance, higher
Several investigators have demonstrated athero- lipoprotein(a) level, proteinuria, azathioprine treat-
sclerotic heart disease in asymptomatic patients with ment, and prednisone dose. In multivariable modeling,
SLE across the age range, using noninvasive imaging both azathioprine use (P ⴝ 0.005 for the mean-mean
techniques such as computerized tomography, carotid model and P ⴝ 0.102 for the mean-max model) and male
ultrasound, and arterial flow–mediated dilatation (3,5– sex (P < 0.001) were associated with increases in the
8). These studies have identified a variety of factors mean-max CIMT. A moderate dosage of prednisone
contributing to the development of subclinical athero- (0.15–0.4 mg/kg/day) was associated with decreases in
sclerosis, including traditional risk factors (obesity, the mean-max CIMT (P ⴝ 0.024), while high-dose and
smoking, glucose intolerance, family history), medica- low-dose prednisone were associated with increases in
tions, hypertension, increased homocysteine levels, and the mean-mean common CIMT (P ⴝ 0.021) and the
renal disease (9,10), but also indicate that important mean-max CIMT (P ⴝ 0.064), respectively. BMI (P <
additional factors exist in this population. The identity of 0.001) and creatinine clearance (P ⴝ 0.031) remained
lupus-specific factors is not well defined but likely associated with increased mean-mean common CIMT,
relates to the immune pathogenesis of SLE (6,11), while increasing age (P < 0.001) and increasing lipopro-
implicating SLE itself as a potent independent risk tein(a) level (P ⴝ 0.005) were associated with increased
factor. Although few published studies specifically ad- dress cardiovascular risk factors in pediatric SLE, the Conclusion. Traditional as well as nontraditional
pathogenesis of atherosclerosis in children and adoles- risk factors were associated with increased CIMT in this
cents with SLE is likely multifactorial, as it is in adults.
cohort of patients in the APPLE trial. Azathioprine
Investigators have shown that nephrotic-range protein- treatment was associated with increased CIMT. The
uria in children with SLE is associated with carotid relationship between CIMT and prednisone dose may
intima-media thickness (CIMT) (5).
not be linear.
Because children and adolescents with SLE have fewer comorbidities and traditional cardiovascular risk Over the last 3 decades, lupus-related mortality factors compared with adults with lupus, this population has decreased in all categories except cardiovascular presents a unique opportunity to further understand the disease (1). The lack of improvement in cardiovascular role of SLE in the pathogenesis of premature athero- morbidity and mortality in systemic lupus erythematosus sclerosis. Currently, no standard imaging procedures are (SLE) may reflect improved survival, as patients with available for the assessment of preclinical atherosclero- lupus now live long enough to experience the develop- sis in young individuals; however, the use of noninvasive ment of cardiovascular disease. However, women with techniques in ongoing observational studies and clinical SLE who are younger than age 40 years are at high risk trials for a variety of disorders may result in clinically for the development of myocardial infarction, with the useful technology in the future.
risk being up to 50-fold higher than that in control CIMT, as measured by carotid ultrasonography, populations (2). The unique convergence of immune is a well-studied surrogate marker of atherosclerosis that and vascular pathology in SLE places young people at has been shown to predict cardiovascular events (stroke unusually high risk of cardiovascular disease (3); this risk and myocardial infarction) (12). It has been used widely is not well addressed by current treatment or prevention as an outcome measure in clinical trials, including studies in children who have an increased risk of athero- Atherosclerosis begins in childhood, even in the sclerosis due to familial hypercholesterolemia (13,14).
absence of SLE. Autopsy studies of trauma victims Although CIMT increases with age, the progression rate commonly showed fatty streaks in the arteries of chil- in healthy children and adolescents is negligible (0.000 dren (4). Children and adolescents with SLE are partic- mm/year between the ages of 10 years and 24 years in ularly at risk as they age, given their life-long burden of females and between the ages of 10 years and 19 years in exposure to multisystem inflammatory disease with a males) (15). In contrast, the progression rate in adult high atherogenic potential. As a result, the prevention of women with SLE is 0.004 mm/year in those ages 30 years and younger and 0.010 mm/year in those older than age Inclusion and exclusion criteria for the APPLE trial* 30 years (Manzi S: personal communication) (16). Pop- Inclusion criteria ulation studies show that increases in carotid wall thick- SLE diagnosed based on 1997 revised ACR criteria ening as small as 0.1 mm strongly correlate with in- Weight ⱖ25 kgOutpatient status creases in the incidence of coronary and cerebral Age 10 years to 21 years vascular events (17).
Ability to complete self-report questionnaires in English or The ongoing Atherosclerosis Prevention in Pedi- Willingness to comply with recommended diet atric Lupus Erythematosus (APPLE) trial is designed to Willingness to use approved birth control method prospectively assess the effect of atorvastatin on the Exclusion criteria progression of CIMT in children and young adults with Drug-induced lupusLiver disease (ALT or AST level of ⬎2-fold the upper limit of SLE. Although the final results of the APPLE trial will not be available until 2010, the purpose of the current Myositis (creatine kinase level of ⬎3-fold the upper limit of analysis is to investigate the association of traditional Nephrotic syndrome (urinary protein excretion rate of ⬎3 gm/24 and nontraditional risk factors with a surrogate marker hours or protein:creatinine ratio of ⬎3.0 and serum albumin of early atherosclerosis (CIMT) in children and adoles- level of ⬍2.3 gm/dl) cents with SLE, using baseline data from the APPLE Hypercholesterolemia (total cholesterol level of ⬎350 mg/dl)Presence of xanthoma trial. The APPLE Investigators are listed in Appendix A.
Familial hypercholesterolemiaRenal insufficiency (dialysis or creatinine level of ⬎2.5 mg/dl)Current use of cyclosporine or tacrolimus PATIENTS AND METHODS
Pregnant or lactatingCurrent use of oral contraceptives containing ⱖ50 ␮g estradiol Participants. A prospective multicenter cohort of 221
Life-threatening non-SLE illness children and adolescents with SLE from 21 sites in North Current drug or alcohol abuse America underwent ultrasound measurement of CIMT at the Inability to obtain adequate-quality CIMT images time of enrollment, as part of the APPLE trial. The APPLE * APPLE ⫽ Atherosclerosis Prevention in Pediatric Lupus Erythem- trial is a double-blind, randomized, placebo-controlled study atosus; SLE ⫽ systemic lupus erythematosus; ACR ⫽ American designed to determine the efficacy and safety of atorvastatin in College of Rheumatology; ALT ⫽ alanine aminotransferase; AST ⫽ preventing the progression of atherosclerosis in children and aspartate aminotransferase; CIMT ⫽ carotid intima-media thickness.
adolescents with SLE treated for 3 years. The primary endpoint is the rate of progression of the mean IMT in thecommon carotid artery over 3 years. Subjects were randomizedto receive either placebo or atorvastatin (10 mg or 20 mg daily,depending on weight); the study drug was provided by Pfizer.
Race and ethnicity were self-reported. For the purposes of this All participants are treated with folate, hydroxychloroquine, analysis, participants were categorized as having minority and low-dose aspirin and receive dietary recommendations status if they were either Hispanic or nonwhite.
and risk factor counseling. Enrollment was completed in 2006, SLE disease activity and damage were assessed using and the study results are expected in 2010.
the modified Safety of Estrogens in Lupus Erythematosus: All participants met the American College of Rheu- National Assessment version of the SLE Disease Activity matology (ACR) revised criteria for SLE as updated in 1997 Index (SLEDAI) (19) and the Systemic Lupus International (18). Additional inclusion requirements included weight ⱖ25 Collaborating Clinics/ACR Damage Index (SDI) (20). For kg, outpatient status, age 10 years to 21 years, the ability to these analyses, the total SLEDAI score was used as a contin- complete self-report questionnaires in English or Spanish, uous variable. SDI scores were analyzed as a dichotomous willingness to comply with the recommended diet, and willing- variable, with patients grouped according to a score of 0 (no ness to use an approved birth control method. Patients were damage) or a score of ⱖ1 (damage present). Additionally, excluded from the study if they had active nephrotic syndrome, SLE-specific laboratory testing was performed. Proteinuria myositis, liver disease, renal insufficiency, or hypercholesterol- was defined as a spot urine protein:creatinine ratio of ⬎0.5 or emia warranting treatment (total cholesterol level ⬎350 mg/dl) a timed urinary protein excretion rate of ⬎500 mg/24 hours.
at baseline. Table 1 shows the full inclusion and exclusion Creatinine clearance was calculated using the formula de- criteria for the APPLE trial.
scribed by Schwartz et al (21).
Clinical variables of interest. The APPLE trial base-
Traditional risk factors for atherosclerosis, including line assessment included demographics, history and physical hypertension, family history of cardiovascular disease, smok- examination performed by a pediatric rheumatologist, chart ing, diet, exercise, and body mass index (BMI), were assessed.
review, and patient and parent questionnaires. Variables of A history of hypertension referred to reported current or interest for this analysis were SLE clinical and laboratory historic hypertension. Similarly, a history of nephrotic syn- parameters, non–lupus-related and lupus-related medical his- drome or nephritis referred to a reported current or historic tory, family history, current medications, traditional risk fac- event. A family history of cardiovascular disease was defined as tors for atherosclerosis, and SLE disease activity and damage.
a participant-reported cerebrovascular accident, myocardial PREMATURE ATHEROSCLEROSIS IN PEDIATRIC SLE infarction, angina, or atherosclerosis in a parent or grand- far walls of the right and left common carotid arteries, carotid bifurcation, and internal carotid artery. For each of the 4 Oral glucocorticoid dosages were weight-adjusted and measurement sites in the common carotid artery, a mean recorded as prednisone equivalents (mg/kg/day). Processing to CIMT value, defined as the average of the 4 angle-specific determine 12-hour fasting lipid, homocysteine, lipoprotein(a) mean CIMT values, was also calculated. The 4 mean CIMT (Lp[a]), and high-sensitivity C-reactive protein (hsCRP) levels values were then averaged to determine the mean-mean was performed centrally (PPD, Highland Heights, KY). Total common CIMT. Per protocol, duplicate ultrasound examina- cholesterol, triglyceride, and high-density lipoprotein levels tions were performed within a 4-week interval, at baseline and were measured directly, and the low-density lipoprotein (LDL) prior to randomization, for each participant. The values for the level was calculated using the equation described by Friedwald mean-mean common CIMT and the mean-max CIMT were et al (22). All other laboratory assessments were performed averaged for these 2 examinations to obtain stable baseline locally. HsCRP and Lp(a) values were log transformed to measures of CIMT. Both CIMT summary measures are recog- achieve approximately normal distributions.
nized as noninvasive surrogate predictors of future cardiovascular Measurement of CIMT. Two CIMT measurements
events and are widely used in clinical trials testing the efficacy of were obtained at baseline, using a standardized ultrasonog- drugs in slowing or reversing atherosclerosis (26–29).
raphy protocol similar to those used in several previous clinical Statistical analysis. The characteristics of the study
trials (23–25). Carotid ultrasound examinations were per- sample were summarized using descriptive statistics, withdichotomous or ordinal data presented as percentages and formed using portable Siemens Cypress systems (Siemens, continuous data presented as means, SDs, and medians. Dif- Munich, Germany), which were shared between clinical centers.
ferences between groups were assessed with either the chi- All scanners were equipped with 7L3 transducers, a cardiology square test or the nonparametric Wilcoxon's test.
package with electrocardiogram (EKG) tracing, and specific Using cross-sectional baseline data from the APPLE APPLE presets in order to reduce variability across centers.
trial, we first investigated potential risk factors and their Standardized longitudinal B-mode images were col- univariable relationship with each of the 2 baseline CIMT lected for 3 arterial segments, defined relative to the tip of the outcomes: 1) mean-mean common CIMT (in millimeters) and flow divider as the common carotid artery (10–20 mm proximal 2) mean-max CIMT (in millimeters). The potential risk factors to the tip of the flow divider), the carotid bifurcation (from the included demographics, physical characteristics, renal func- tip of the flow divider to 10 mm proximal to the tip of the flow tion, markers of inflammation, medical history, lipid levels, divider), and the proximal 10 mm of the internal carotid artery.
current or recent medications, and disease indices. Scatterplots Near and far walls were imaged simultaneously in the common with overlay of locally weighted scatterplot smoothing carotid artery but separately in the carotid bifurcation and (LOESS) curves (30) were produced to examine the descrip- internal carotid artery, to improve the ability to align each wall tive relationship between continuous risk factors and CIMT horizontally in these segments. For each arterial segment, outcome variables. Histograms of CIMT values by levels of the images were selected at 90, 120, 150, and 180 degrees, as categorical risk factors were used to examine the relationship marked on the Meijer Arc, when the right side was scanned between categorical variables and CIMT outcomes. Linear and at 270, 240, 210, and 180 degrees when the left side was regression models were fitted using 1 potential risk factor for CIMT at a time to select candidate variables for multivariable Image selections were saved as 5-second digital clips modeling. Residual plots of the regression models were used to and written to 640-MB magnetic optical disks for transfer to a assess the suitability of the linear model. For nonlinear rela- central reading center (Ward A. Riley Ultrasound Center, tionships between the continuous risk factors and the CIMT Wake Forest University School of Medicine, Winston-Salem, outcome, transformations or piecewise linear regression models NC). All ultrasound scans were read by a single experienced were used to better characterize the univariable relationship.
reader, using Image-Pro software (Media Cybernetics, Be- Variables with a P value of ⱕ0.2 in the univariable thesda, MD). For each image sequence, the single reader model were included in multivariable linear regression model- selected 1 frame for measurement when the heart was in ing. Several model-building procedures were used, including systole (when the EKG tracing is on the QRS complex).
forward, backward, and stepwise regression, to produce 2 final Leading (far wall) and trailing (near wall) edges of visualized models, 1 for each of the 2 CIMT outcomes. Several methods blood–intima and media–adventitia boundaries were traced were used to assess the colinearity and stability of the final with a computer mouse–controlled caliper within a region of models, including assessment of the association between inde- interest specified by the reader. Quality assurance procedures pendent variables, colinearity diagnostics within the regres- included central training and certification of all sonographers sion framework, and use of Mallows' Cp values. All statistical and the reader as well as regular site visits and performance inference was done within a linear regression model framework, and a P value of ⱕ0.1 is considered to be statistically significant The combination of 3 arterial segments, 2 walls, and 2 due to the exploratory nature of the analyses. All calculations sides of the neck provides a set of 12 CIMT measurement sites, were performed using SAS version 8.2 software (31).
each of which was imaged from 4 angles. For each measure-ment site, a maximum CIMT value, defined as the largest of the 4 angle-specific maximum CIMT values, was calculated.
The 12 maximum CIMT values were then averaged to deter- Patient characteristics. A summary of the base-
mine the mean maximal (mean-max) CIMT over the near and line demographic and clinical characteristics of the Baseline characteristics of the APPLE subjects* Age, years (n ⫽ 221) 15.7 ⫾ 2.6 (15.5) Female sex, no. (%) (n ⫽ 221) Weight, kg (n ⫽ 221) 62.0 ⫾ 17.2 (58.7) Body mass index, kg/m2 (n ⫽ 221) 24.4 ⫾ 5.3 (23.4) Duration of lupus, months (n ⫽ 220) 31.2 ⫾ 28.5 (25.0) Hispanic ethnicity, no. (%) (n ⫽ 221) Race, no. (%) (n ⫽ 221) Minority status (Hispanic or nonwhite), no. (%) (n ⫽ 221) Self-reported history of smoking, no. (%) (n ⫽ 221) Prednisone dosage, mg/kg/day (n ⫽ 218) 0.19 ⫾ 0.19 (0.15) Family history of cardiovascular disease, no. (%) (n ⫽ 211) Creatinine clearance, ml/minute/m2 (n ⫽ 216) 139.4 ⫾ 33.0 (134.1) Proteinuria, no. (%) (n ⫽ 220) History of renal abnormalities, no. (%) Hypertension (n ⫽ 214) Nephrotic syndrome (n ⫽ 219) Nephritis (n ⫽ 219) Other (n ⫽ 190) Current medications, no. (%) (n ⫽ 221) Angiotensin-converting enzyme inhibitors Nonsteroidal antiinflammatory drugs Acetylsalicylic acid Total cholesterol, mg/dl (n ⫽ 211) 155.1 ⫾ 38.0 (148.0) HDL cholesterol, mg/dl (n ⫽ 211) 46.3 ⫾ 12.8 (44.0) LDL cholesterol, mg/dl (n ⫽ 210) 86.4 ⫾ 31.4 (78.5) Triglycerides, mg/dl (n ⫽ 211) 114.0 ⫾ 66.4 (101.0) Homocysteine, ␮moles/liter (n ⫽ 207) 7.5 ⫾ 3.1 (6.8) Lp(a), mg/dl (n ⫽ 206) 23.1 ⫾ 26.8 (13.5) Log(Lp[a]) (n ⫽ 206) 2.5 ⫾ 1.2 (2.6) Mean-max CIMT, mm (n ⫽ 221) 0.585 ⫾ 0.052 (0.579) Mean-mean common CIMT, mm (n ⫽ 221) 0.468 ⫾ 0.041 (0.466) HsCRP, mg/liter (n ⫽ 202) 3.6 ⫾ 13.9 (0.7) Log(hsCRP) (n ⫽ 202) ⫺0.27 ⫾ 1.52 (⫺0.36) Erythrocyte sedimentation rate, mm/hour (n ⫽ 148) 23.1 ⫾ 22.9 (16.0) Anti-dsDNA antibody positive, no. (%) (n ⫽ 221) SLEDAI total score (n ⫽ 221) 4.6 ⫾ 4.2 (4.0) SDI score ⬎0, no. (%) (n ⫽ 221) * Creatinine clearance was calculated using the formula described by Schwartz et al (21) (see Patients andMethods). Proteinuria was defined as a urinary protein excretion rate of ⬎500 mg/24 hours or a spotprotein: creatinine ratio of 0.5. Some subjects reported more than 1 race. Except where indicatedotherwise, values are the mean ⫾ SD (median). APPLE ⫽ Atherosclerosis Prevention in Pediatric LupusErythematosus; HDL ⫽ high-density lipoprotein; LDL ⫽ low-density lipoprotein; Lp(a) ⫽ lipoprotein(a);max ⫽ maximal; CIMT ⫽ carotid intima-media thickness; hsCRP ⫽ high-sensitivity C-reactive protein;anti-dsDNA ⫽ anti–double-stranded DNA; SLEDAI ⫽ Systemic Lupus Erythematosus Disease ActivityIndex; SDI ⫽ Systemic Lupus International Collaborating Clinics/American College of RheumatologyDamage Index.
participants is shown in Table 2. Among the 221 enrolled (Hispanic or nonwhite). Patients had a mean disease patients, the mean age was 15.7 years (range 10.1–21.7 duration of 31 months (median 25, interquartile range years), 83% were female, and 65% had minority status [IQR] 8–46) and had mildly active disease, with an PREMATURE ATHEROSCLEROSIS IN PEDIATRIC SLE Univariable relationships between clinical risk factors and CIMT* Mean-mean common CIMT, mm Mean-max CIMT, mm SLEDAI, total score ⫺0.0002 (0.0008) Anti-dsDNA antibody positive Duration of lupus, months Creatinine clearance, ml/minute/m2 History of nephrotic syndrome/ History of hypertension Family history of cardiovascular Low-density lipoprotein cholesterol ⫺0.00006 (0.00009) ⫺0.00007 (0.0001) High-density lipoprotein cholesterol ⫺0.0002 (0.0002) Minority race (Hispanic or nonwhite) NSAIDs, current use Azathioprine, current use Methotrexate, current use Acetylsalicylic acid, current use ACE inhibitor, current use Prednisone dosage ⬍0.15 mg/kg/day Prednisone dosage 0.15–0.40 mg/kg/day Prednisone dosage ⬎0.40 mg/kg/day * Creatinine clearance was calculated using the formula described by Schwartz et al (21) (see Patients and Methods). Proteinuria was defined as aurinary protein excretion rate of ⬎500 mg/24 hours or a spot protein:creatinine ratio of 0.5. CIMT ⫽ carotid intima-media thickness; max ⫽maximal; hsCRP ⫽ high-sensitivity C-reactive protein; SLEDAI ⫽ Systemic Lupus Erythematosus Disease Activity Index; anti-dsDNA ⫽anti–double-stranded DNA; Lp(a) ⫽ lipoprotein(a); SDI ⫽ Systemic Lupus International Collaborating Clinics/American College of RheumatologyDamage Index; NSAIDs ⫽ nonsteroidal antiinflammatory drugs; ACE ⫽ angiotensin-converting enzyme.
average SLEDAI score of 4.6. Most patients had no BMI, male sex, higher creatinine clearance, higher Lp(a) damage as measured by the SDI. Of the 178 patients level, presence of proteinuria, current azathioprine use, receiving prednisone at baseline, the mean weight- and current weight-adjusted prednisone dose. The com- adjusted dosage was 0.19 mg/kg/day (median 0.15, IQR plete univariable modeling results are shown in Table 3.
0.05–0.26). The 2 baseline (prerandomization) CIMT The relationship between weight-adjusted pred- assessments were performed a median of 9 days apart nisone dose and both the mean-mean common and (IQR 2–21).
the mean-max CIMT was found to be nonmonotonic.
Univariable analysis. The average of the 2 base-
We identified 3 different slopes upon examining the line CIMT measurements rather than a single baseline relationship between prednisone and CIMT and there- measurement was used for analysis, in order to increase fore performed 3 piecewise linear regressions using the precision. The 2 baseline scans were strongly correlated, following weight-adjusted prednisone dosage ranges: with Pearson's correlation coefficients of 0.75 for the 0.0–0.15 mg/kg/day, 0.15–0.4 mg/kg/day, and ⬎0.4 mg/ mean-mean common CIMT and 0.76 for the mean-max kg/day. Descriptive relationships are shown in Figure 1, CIMT. Based on either the mean-mean common or with LOESS curves overlaying the scatterplots of mean- mean-max CIMT as the dependent variable, univariable mean common or mean-max CIMT versus weight- analysis suggested the following risk factors to be signif- adjusted prednisone dosage.
icantly associated with increased CIMT: increasing Multivariable analysis. Different model-building
age, longer duration of SLE, minority status, higher procedures yielded the same 2 final models. The results other variables. Both azathioprine use (P ⫽ 0.005 for themean-mean model and P ⫽ 0.102 for the mean-maxmodel) and male sex (P ⬍ 0.001) were associated withincreases in the mean-max CIMT. Moderate-dosageprednisone (0.15–0.4 mg/kg/day) was associated withdecreases only in the mean-max CIMT (P ⫽ 0.024),while high-dosage (⬎0.4 mg/kg/day) and low-dosage(⬍0.15 mg/kg/day) prednisone were associated with themean-mean common CIMT (P ⫽ 0.021) and the mean-max CIMT (P ⫽ 0.064), respectively. BMI (P ⬍ 0.001)and creatinine clearance (P ⫽ 0.031), presumably re-lated to hyperfiltration (32), remained associated onlywith increased mean-mean common CIMT, while in-creasing age (P ⬍ 0.001) and increasing Lp(a) levels(P ⫽ 0.005) were associated only with increased mean-max CIMT in the multivariable models. For the mean-mean common CIMT model, R2 ⫽ 0.22, and for themean-max model, R2 ⫽ 0.20. Thus, the potential riskfactors explained ⬃21% of variation in the CIMT vari-ables. The multivariable models were stable, and noimportant colinearity of variables was noted using re-gression diagnostics, including Mallows' Cp values.
Analysis of the baseline data from the APPLE trial, a racially and ethnically diverse multicenter pro-spective cohort of patients with pediatric SLE, revealsthat both traditional and nontraditional risk factors areassociated with increased CIMT in children with SLE.
Traditional risk factors predictive of increased CIMT inthis cohort include increased BMI, male sex, and in-creasing age. Perhaps more intriguing is the identifica-tion of nontraditional risk factors, including treatmentwith azathioprine and the weight-adjusted prednisonedosage.
Figure 1. A, Locally weighted scatterplot smoothing (LOESS) curve
The frequent and long-term use of glucocorti- showing the mean-mean common carotid intima-media thickness
(CIMT) versus the weight-adjusted prednisone dosage. B, LOESS
coids in children and adolescents with SLE likely plays a curve showing the mean maximal (mean-max) CIMT versus the role in the prevalence of dyslipidemia and atherosclero- weight-adjusted prednisone dosage. Broken lines represent the 95% sis, but this has not been well characterized to date.
Some studies in adult SLE suggest an association be-tween prednisone use and both increased total choles-terol levels and CIMT (33,34). In contrast, other studies of the multivariable analysis for each mean-mean com- in adults with SLE suggest that those with carotid plaque mon CIMT and mean-max CIMT are shown in Table 4, received less aggressive immunosuppressive therapy (in- including the adjusted slope estimates from the final cluding less prednisone and cyclophosphamide) than linear regression models of the potential risk factors with those without plaque, suggesting a potentially protective the CIMT variable, corresponding standard errors, and effect of immunosuppressive therapy on the develop- P values. Duration of SLE, minority status, and presence ment of atherosclerosis, possibly due to better control of of proteinuria are not present in the final models due to underlying inflammatory disease (6,10,35). In the cur- nonsignificant associations after adjustment for the rent study, moderate-dose prednisone was associated PREMATURE ATHEROSCLEROSIS IN PEDIATRIC SLE Multivariable adjusted relationships between clinical risk factors and CIMT* Mean-max CIMT, mm Creatinine clearance, ml/minute/m2 Azathioprine, current use Prednisone dosage 0.00–0.15 mg/kg/day Prednisone dosage 0.15–0.40 mg/kg/day Prednisone dosage ⬎0.40 mg/kg/day * Only associations with a P value less than or equal to 0.1 were included in the final models. CIMT ⫽carotid intima-media thickness; max ⫽ maximal; NI ⫽ not included in the final models due tononsignificant associations; Lp(a) ⫽ lipoprotein(a).
with decreased mean-max CIMT, while higher-dose or including SLE disease activity and a history of nephritis, lower-dose prednisone may predict increased mean- and is not explained by colinearity with other variables.
mean common or mean-max CIMT, respectively. The To further understand the relationship between demonstration of different relationships between pred- azathioprine and CIMT, the 30 patients receiving aza- nisone dose and CIMT may suggest that steroids are thioprine were compared with the 191 patients who were affecting more than 1 pathway in atherogenesis. At high not receiving azathioprine. As shown in Table 5, patients doses, prednisone may increase CIMT by increasing the receiving azathioprine were, on average, 1 year older levels of traditional risk factors, such as total cholesterol, (mean age 16.7 years versus 15.6 years; P ⫽ 0.027), had LDL cholesterol, and BMI. In contrast, with the current a slightly longer disease duration (median duration 31.5 understanding of the importance of inflammatory mech- months versus 21.5 months; P ⫽ 0.005), and were more anisms in the pathogenesis of atherosclerosis (36), likely to have an SDI score of ⬎0 (43.3% versus 24.1%; lower-dose prednisone may be associated with higher P ⫽ 0.027). Upon removing azathioprine from multiva- CIMT due to the presence of ongoing active inflamma- riable modeling, the duration of SLE became statistically tory disease. Given the cross-sectional nature of the significant, but the SDI score did not, suggesting a data, it was not possible to assess the association be- possible confounding effect due to the relationship tween the cumulative steroid dose and CIMT. However, between azathioprine use and disease duration. How- this will be assessed in a longitudinal analysis in the ever, when both variables (azathioprine and disease APPLE trial. The differences in CIMT based on steroid duration) were included in the multivariable modeling, dose may provide insight into previous conflicting results only azathioprine remained significant, suggesting that concerning the impact of oral prednisone on cardiovas- azathioprine has a stronger association with CIMT. In cular risk and suggest that there may be a "therapeutic contrast, age was a significant risk factor for greater window" concerning cardiovascular toxicity in this pop- CIMT, even when azathioprine was included in the multivariable model for mean-max CIMT, suggesting Other medications used in the treatment of SLE that both age and azathioprine have independent effects may impact the development of atherosclerosis. For on CIMT. Given the relative affordability of azathio- example, cyclosporine contributes to dyslipidemia prine, additional analyses to explore whether azathio- (37,38), while hydroxychloroquine improves lipid pro- prine use reflected socioeconomic status revealed no files and is associated with a decreased incidence of correlation between the use of azathioprine and house- cardiovascular events (39). Azathioprine, mycopheno- hold income or parental education (data not shown).
late mofetil, and cyclophosphamide have not previously Although the potential proatherogenic effects of been shown to have a significant impact on lipid profiles azathioprine are not known, an interesting in vitro or atherogenesis (39). A surprising finding in this study observation using human umbilical vein endothelial cells was that azathioprine, unlike the other immunosuppres- showed that azathioprine and its active metabolite, sant agents, was associated with increased CIMT. This 6-mercaptopurine, caused dose-dependent decreases in effect appears to be independent of other variables, endothelial cell proliferation and altered endothelial Comparison of patients receiving azathioprine and those not receiving azathioprine* Age, mean ⫾ SD years (median) 15.6 ⫾ 2.7 (15.5) 16.7 ⫾ 2.4 (17.1) Weight, mean ⫾ SD kg (median) 61.7 ⫾ 16.9 (59.5) 63.9 ⫾ 19.0 (58.0) Body mass index, mean ⫾ SD kg/m2 (median) 24.3 ⫾ 5.3 (23.3) 24.9 ⫾ 5.8 (24.1) Duration of lupus, mean ⫾ SD months (median) 28.7 ⫾ 25.9 (21.5) 47.0 ⫾ 38.3 (31.5) Minority status (Hispanic ethnicity or nonwhite race) Family history of cardiovascular disease History of smoking Prednisone dosage, mean ⫾ SD mg/kg/day (median) 0.19 ⫾ 0.19 (0.14) 0.18 ⫾ 0.17 (0.15) Creatinine clearance, mean ⫾ SD ml/minute/m2 (median) 140.0 ⫾ 34.0 (134.4) 134.9 ⫾ 25.4 (133.2) Renal abnormalities Current medications Angiotensin-converting enzyme inhibitor Acetylsalicylic acid Nonsteroidal antiinflammatory drugs Anti-dsDNA antibody positive (reported in SLEDAI) SLEDAI, total score 4.6 ⫾ 4.2 (4.0) 4.7 ⫾ 4.6 (4.0) Erythrocyte sedimentation rate, mm/hour 22.8 ⫾ 23.3 (15.0) 24.9 ⫾ 20.9 (18.0) Total cholesterol, mean ⫾ SD mg/dl (median) 155.1 ⫾ 38.4 (148.0) 155.6 ⫾ 35.4 (153.0) Triglycerides, mean ⫾ SD mg/dl (median) 115.3 ⫾ 69.3 (102.0) 106.1 ⫾ 44.9 (97.0) HDL cholesterol, mean ⫾ SD mg/dl (median) 46.0 ⫾ 12.4 (44.0) 48.2 ⫾ 14.8 (49.0) LDL cholesterol, mean ⫾ SD mg/dl (median) 86.4 ⫾ 31.3 (79.0) 86.2 ⫾ 32.3 (78.0) Homocysteine, mean ⫾ SD ␮moles/liter (median) 7.4 ⫾ 3.2 (6.8) 7.6 ⫾ 2.8 (6.5) Lp(a), mean ⫾ SD mg/dl (median) 23.1 ⫾ 27.6 (14.0) 23.0 ⫾ 21.8 (13.0) Log(Lp[a]), mean ⫾ SD (median) 2.5 ⫾ 1.2 (2.6) 2.6 ⫾ 1.1 (2.6) HsCRP, mean ⫾ SD mg/liter (median) 3.35 ⫾ 14.40 (0.60) 4.88 ⫾ 10.83 (0.80) Log(hsCRP), mean ⫾ SD (median) ⫺0.34 ⫾ 1.47 (⫺0.51) 0.11 ⫾ 1.77 (⫺0.22) Mean-max CIMT, mean ⫾ SD mm (median) 0.466 ⫾ 0.040 (0.462) 0.483 ⫾ 0.041 (0.475) Mean-mean common CIMT, mean ⫾ SD mm (median) 0.582 ⫾ 0.050 (0.578) 0.605 ⫾ 0.058 (0.611) * Except where indicated otherwise, values are the number of patients/number of patients assessed (%). Creatinine clearance was calculated usingthe formula described by Schwartz et al (21) (see Patients and Methods). Proteinuria was defined as a urinary protein excretion rate of ⬎500 mg/24hours or a spot protein:creatinine ratio of 0.5. SDI ⫽ Systemic Lupus International Collaborating Clinics/American College of RheumatologyDamage Index; anti-dsDNA ⫽ anti–double-stranded DNA; SLEDAI ⫽ Systemic Lupus Erythematosus Disease Activity Index; HDL ⫽ high-densitylipoprotein; LDL ⫽ low-density lipoprotein; Lp(a) ⫽ lipoprotein(a); hsCRP ⫽ high-sensitivity C-reactive protein; max ⫽ maximal; CIMT ⫽ carotidintima-media thickness.
nucleotide balance by reducing the levels of intracellular (LUMINA) cohort (42). Additional study is necessary to ATP and GTP (40). It is possible that azathioprine has a further explore this potentially significant association.
direct deleterious impact on endothelial cell function Creatinine clearance may be increased in settings and the capacity for endothelial repair in SLE. In of glomerular hyperfiltration, such as proteinuria or addition, in vitro studies suggest that azathioprine may hypertension (32,43). Although patients with significant increase homocysteine levels; however, we did not ob- renal insufficiency or nephrotic syndrome were excluded serve a significant relationship between CIMT and ho- from the APPLE trial, many participants had protein- mocysteine levels in this analysis (41). It is interesting to uria and hypertension. With univariable analysis, both note that azathioprine was demonstrated to be a con- proteinuria and creatinine clearance were associated tributing predictor of vascular events in a racially/ with CIMT, but only creatinine clearance remained ethnically diverse cohort of adults with SLE in the significant in the multivariable model. It is possible that multicenter LUpus in MInorities, NAture versus nurture creatinine clearance and proteinuria are correlated; PREMATURE ATHEROSCLEROSIS IN PEDIATRIC SLE however, longitudinal study of this cohort will be neces- firmed by future studies. In addition, as in any clinical sary to further evaluate this effect.
trial, there is an inherent selection bias such that the In this study, CIMT was assessed using 2 different APPLE cohort may not be completely representative of approaches: the mean-mean common CIMT and the pediatric lupus patients as a whole. For example, obese mean-max CIMT. The correlation between the 2 out- patients were more likely to be excluded due to the comes was 0.78, suggesting that the 2 outcomes are inability to obtain high-quality CIMT images. Patients correlated but are not perfectly linearly related. The 2 with nephrotic-range proteinuria, very elevated choles- approaches differ in terms of the segments of the carotid terol levels, or renal insufficiency, all of whom were also artery measured as well as how the thickness is assessed excluded from the APPLE study, are known to have (average versus maximal thickness of the measured increased CIMT (5). These subsets of pediatric SLE segments). Both approaches have been used in prior patients may have increased CIMT and increased lipid clinical trials, and a review of the literature reveals no abnormalities and different relationships between the clear superiority of one method over the other in risk factors and CIMT. Thus, their exclusion may affect predicting atherosclerotic risk (44,45). In multiple re- the generalizability of the reported findings.
gression models, we identified factors (sex and medica- The presented models explain only ⬃21% of the tion use) that were significantly associated with both variation in CIMT, suggesting the influence of other outcomes and others that were significantly associated measured or unmeasured risk factors that were not with only 1 outcome (age and Lp[a] were related to the identified in the current study. Cross-sectional analyses mean-max CIMT only, while BMI and creatinine clear- of the baseline data also limit the ability to draw ance were related to the mean-mean common CIMT only).
conclusions about causation between risk factors and Segment-specific risk factor associations have CIMT outcomes. Once the APPLE trial is complete, been reported previously in IMT studies and are gener- longitudinal analysis will provide more meaningful as- ally considered to reflect the influence on focal thicken- sessment of the impact of glucocorticoids, azathioprine, ing of local phenomena such as sheer stress and arterial and other factors on CIMT. Unfortunately, the APPLE remodeling (46,47). Interestingly, Espeland et al (48) study does not include matched normal control subjects.
reported segment-specific risk factor associations in Because of the inherent variation in CIMT measure- patients with coronary artery disease but not in control ment techniques, the absolute thickness cannot be com- subjects of similar age. The observation that obesity is pared between studies. Therefore, there is no potential more closely related to common CIMT than to bifurca- to compare the CIMT results in this cohort with those in tion or internal CIMT is consistent with prior studies inadults (47,48). Associations between estimated creati- a healthy population.
nine clearance and common CIMT have previously been In summary, traditional as well as nontraditional reported in children with chronic renal failure (49) and risk factors predict increased CIMT in pediatric patients in young adults with myocardial infarction (50). Studies with SLE. Of interest, treatment with azathioprine was of children and adolescents with growth hormone defi- associated with increased CIMT, and the relationship ciency (51) and adult women with SLE and cardiovas- with the weight-adjusted oral prednisone dose may not cular disease (52) have also documented significant be linear. Future longitudinal analysis of data from the increases in both the Lp(a) level and common CIMT APPLE trial will further clarify the role of medications relative to normal control subjects. Segment-specific such as azathioprine and corticosteroids, as well as other differences in associations between CIMT and either potential risk factors for CIMT progression. Under- creatinine clearance or Lp(a) have not been reported standing risk factors in the development of atheroscle- rosis in children and adolescents with SLE will inform This study has several limitations. The sample evidence-based treatment recommendations, to avoid size was powered to detect only differences in the irreversible cardiovascular damage.
primary outcome of the APPLE trial—the progressionrate of CIMT. Thus, it may not have sufficient power to identify all meaningful clinical associations of risk fac- Dr. Schanberg had full access to all of the data in the study tors with baseline CIMT. However, this is the largest and takes responsibility for the integrity of the data and the accuracy cohort of pediatric SLE patients ever examined for of the data analysis.
Study design. Schanberg, Sandborg, Barnhart, Ardoin, Evans.
cardiovascular risk, and the potential relationships re- Acquisition of data. Schanberg, Sandborg, Evans, Mieszkalski, Ilowite,
ported here are hypothesis-generating and can be con- Eberhard, Levy, Kimura, von Scheven, Silverman, Bowyer, Punaro, Singer, Sherry, McCurdy, Klein-Gitelman, Wallace, Silver, Wagner- der Graaf A, de Groot E, et al. Statin treatment in children with Weiner, Higgins, Brunner, Jung, Soep.
familial hypercholesterolemia: the younger, the better. Circulation Analysis and interpretation of data. Schanberg, Sandborg, Barnhart,
Ardoin, Yow, Evans, Mieszkalski, Ilowite, von Scheven, Brunner.
15. Sass C, Herbeth B, Chapet O, Siest G, Visvikis S, Zannad F.
Manuscript preparation. Schanberg, Sandborg, Barnhart, Ardoin,
Intima-media thickness and diameter of carotid and femoral Yow, Evans, Mieszkalski, Ilowite, Eberhard, Levy, Kimura, von arteries in children, adolescents and adults from the Stanislas Scheven, Silverman, Bowyer, Punaro, Singer, Sherry, McCurdy, Klein- cohort: effect of age, sex, anthropometry and blood pressure.
Gitelman, Wallace, Silver, Higgins, Brunner, Jung, Soep, Reed.
J Hypertens 1998;16:1593–602.
Statistical analysis. Schanberg, Sandborg, Barnhart, Ardoin, Yow.
16. Selzer F, Sutton-Tyrrell K, Fitzgerald SG, Pratt JE, Tracy RP, Trial management. Schanberg, Sandborg, Mieszkalski.
Kuller LH, et al. Comparison of risk factors for vascular disease inthe carotid artery and aorta in women with systemic lupuserythematosus. Arthritis Rheum 2004;50:151–9.
17. Lorenz MW, Markus HS, Bots ML, Rosvall M, Sitzer M. Predic- tion of clinical cardiovascular events with carotid intima-media 1. Bjornadal L, Yin L, Granath F, Klareskog L, Ekbom A. Cardio- vascular disease a hazard despite improved prognosis in patients thickness: a systematic review and meta-analysis. Circulation 2007; with systemic lupus erythematosus: results from a Swedish popu- lation based study 1964-95. J Rheumatol 2004;31:713–9.
18. Hochberg MC, for the Diagnostic and Therapeutic Criteria Com- 2. Manzi S, Meilahn EN, Rairie JE, Conte CG, Medsger TA Jr, mittee of the American College of Rheumatology. Updating the Jansen-McWilliams L, et al. Age-specific incidence rates of myo- American College of Rheumatology revised criteria for the clas- cardial infarction and angina in women with systemic lupus sification of systemic lupus erythematosus [letter]. Arthritis erythematosus: comparison with the Framingham Study. Am J Rheum 1997;40:1725.
19. Buyon JP, Petri MA, Kim MY, Kalunian KC, Grossman J, Hahn 3. Westerweel PE, Luyten RK, Koomans HA, Derksen RH, Verhaar BH, et al. The effect of combined estrogen and progesterone MC. Premature atherosclerotic cardiovascular disease in systemic hormone replacement therapy on disease activity in systemic lupus lupus erythematosus [review]. Arthritis Rheum 2007;56:1384–96.
erythematosus: a randomized trial. Ann Intern Med 2005;142(12 4. Berenson GS, Wattigney WA, Tracy RE, Newman WP III, Pt 1):953–62.
Srinivasan SR, Webber LS, et al. Atherosclerosis of the aorta and 20. Gladman D, Ginzler E, Goldsmith C, Fortin P, Liang M, Urowitz coronary arteries and cardiovascular risk factors in persons aged 6 M, et al. The development and initial validation of the Systemic to 30 years and studied at necropsy (The Bogalusa Heart Study).
Lupus International Collaborating Clinics/American College of Am J Cardiol 1992;70:851–8.
Rheumatology damage index for systemic lupus erythematosus.
5. Falaschi F, Ravelli A, Martignoni A, Migliavacca D, Sartori M, Arthritis Rheum 1996;39:363–9.
Pistorio A, et al. Nephrotic-range proteinuria, the major risk factor 21. Schwartz GJ, Haycock GB, Edelmann CM Jr, Spitzer A. A simple for early atherosclerosis in juvenile-onset systemic lupus erythem- estimate of glomerular filtration rate in children derived from atosus. Arthritis Rheum 2000;43:1405–9.
body length and plasma creatinine. Pediatrics 1976;58:259–63.
6. Roman MJ, Shanker BA, Davis A, Lockshin MD, Sammaritano L, 22. Friedwald WT, Levy RI, Frederickson DS. Estimation of the Simantov R, et al. Prevalence and correlates of accelerated concentration of low-density lipoprotein cholesterol in plasma, atherosclerosis in systemic lupus erythematosus. N Engl J Med without use of preparative centrifuge. Clin Chem 1972;18:499–502.
23. Bots ML, Evans GW, Riley W, Meijer R, McBride KH, Paskett 7. Jimenez S, Garcia-Criado MA, Tassies D, Reverter JC, Cervera R, ED, et al. The Osteoporosis Prevention and Arterial effects of Gilabert MR, et al. Preclinical vascular disease in systemic lupus tiboLone (OPAL) study: design and baseline characteristics. Con- erythematosus and primary antiphospholipid syndrome. Rheuma- trol Clin Trials 2003;24:752–75.
tology (Oxford) 2005;44:756–61.
24. Crouse JR III, Grobbee DE, O'Leary DH, Bots ML, Evans GW, 8. Wilson WA, Gharavi AE, Koike T, Lockshin MD, Branch DW, Palmer MK, et al. Measuring effects on intima media thickness: an Piette JC, et al. International consensus statement on preliminary evaluation of rosuvastatin in subclinical atherosclerosis: the ration- classification criteria for definite antiphospholipid syndrome: report ale and methodology of the METEOR study. Cardiovasc Drugs of an international workshop. Arthritis Rheum 1999;42:1309–11.
9. Petri M, Perez-Gutthann S, Spence D, Hochberg MC. Risk factors 25. Kastelein JJ, van Leuven SI, Evans GW, Riley WA, Revkin JH, for coronary artery disease in patients with systemic lupus ery- Shear CL, et al. Designs of RADIANCE 1 and 2: carotid thematosus. Am J Med 1992;93:513–9.
ultrasound studies comparing the effects of torcetrapib/atorvasta- 10. Roman MJ, Crow MK, Lockshin MD, Devereux RB, Paget SA, tin with atorvastatin alone on atherosclerosis. Curr Med Res Opin Sammaritano L, et al. Rate and determinants of progression ofatherosclerosis in systemic lupus erythematosus. Arthritis Rheum 26. Bots ML, Evans GW, Riley W, McBride KH, Paskett ED, 11. Esdaile JM, Abrahamowicz M, Grodzicky T, Li Y, Panaritis C, du Helmond FA, et al. The effect of tibolone and continuous com- Berger R, et al. Traditional Framingham risk factors fail to fully bined conjugated equine oestrogens plus medroxyprogesterone account for accelerated atherosclerosis in systemic lupus erythem- acetate on progression of carotid intima-media thickness: the atosus. Arthritis Rheum 2001;44:2331–7.
Osteoporosis Prevention and Arterial effects of tiboLone (OPAL) 12. Greenland P, Abrams J, Aurigemma GP, Bond MG, Clark LT, study. Eur Heart J 2006;27:746–55.
Criqui MH, et al. Prevention Conference V: beyond secondary 27. Crouse JR III, Raichlen JS, Riley WA, Evans GW, Palmer MK, prevention: identifying the high-risk patient for primary preven- O'Leary DH, et al. Effect of rosuvastatin on progression of carotid tion: noninvasive tests of atherosclerotic burden: Writing Group intima-media thickness in low-risk individuals with subclinical III. Circulation 2000;101:E16–22.
atherosclerosis: the METEOR trial. JAMA 2007;297:1344–53.
13. Wiegman A, Hutten BA, de Groot E, Rodenburg J, Bakker HD, 28. Kastelein JJ, van Leuven SI, Burgess L, Evans GW, Kuivenhoven Buller HR, et al. Efficacy and safety of statin therapy in children JA, Barter PJ, et al. Effect of torcetrapib on carotid atherosclerosis with familial hypercholesterolemia: a randomized controlled trial.
in familial hypercholesterolemia. N Engl J Med 2007;356:1620–30.
29. Bots ML, Visseren FL, Evans GW, Riley WA, Revkin JH, Tegeler 14. Rodenburg J, Vissers MN, Wiegman A, van Trotsenburg AS, van CH, et al. Torcetrapib and carotid intima-media thickness in PREMATURE ATHEROSCLEROSIS IN PEDIATRIC SLE mixed dyslipidaemia (RADIANCE 2 study): a randomised, dou- JR III. Associations of risk factors with segment-specific intimal- ble-blind trial. Lancet 2007;370:153–60.
medial thickness of the extracranial carotid artery. Stroke 1999; 30. Cleveland WS. Robust locally weighted regression and smoothing scatterplots. J Am Stat Assoc 1979;74:829–36.
49. Litwin M, Wuhl E, Jourdan C, Trelewicz J, Niemirska A, Fahr K, 31. SAS/STAT user's guide. Cary (NC): SAS Institute; 1999.
et al. Altered morphologic properties of large arteries in children 32. Palatini P, Mormino P, Dorigatti F, Santonastaso M, Mos L, De with chronic renal failure and after renal transplantation. J Am Toni R, et al. Glomerular hyperfiltration predicts the development Soc Nephrol 2005;16:1494–500.
of microalbuminuria in stage 1 hypertension: the HARVEST.
50. Erzen B, Sabovic M, Sebestjen M, Kaplan-Pavlovcic S, Poredos P.
Kidney Int 2006;70:578–84.
Correlation between glomerular filtration and markers of the 33. Doria A, Shoenfeld Y, Wu R, Gambari PF, Puato M, Ghirardello atherosclerotic process in young patients with myocardial infarc- A, et al. Risk factors for subclinical atherosclerosis in a prospective tion. Int J Cardiol 2007. E-pub ahead of print.
cohort of patients with systemic lupus erythematosus. Ann Rheum 51. Szczepaniska Kostro J, Tolwinska J, Urban M, Gardziejczyk M, Glowinska B. Cardiac mass and function, carotid artery intima 34. Petri M, Lakatta C, Magder L, Goldman D. Effect of prednisone media thickness, homocysteine and lipoprotein levels in children and hydroxychloroquine on coronary artery disease risk factors in and adolescents with growth hormone deficiency. J Pediatr Endo- systemic lupus erythematosus: a longitudinal data analysis. Am J crinol Metab 2004;17:1405–13.
52. Svenungsson E, Jensen-Urstad K, Heimburger M, Silveira A, 35. Tyrrell PN, Beyene J, Benseler SM, Sarkissian T, Silverman ED.
Hamsten A, de Faire U, et al. Risk factors for cardiovascular Predictors of lipid abnormalities in children with new-onset sys- disease in systemic lupus erythematosus. Circulation 2001;104: temic lupus erythematosus. J Rheumatol 2007;34:2112–9.
36. Hansson GK. Inflammation, atherosclerosis, and coronary artery disease. N Engl J Med 2005;352:1685–95.
APPENDIX A: THE APPLE INVESTIGATORS
37. Stone NJ. Secondary causes of hyperlipidemia. Med Clin North The APPLE Investigators, in addition to the authors of this 38. Marcen R, Chahin J, Alarcon A, Bravo J. Conversion from article, are as follows: Esi Morgan Dewitt, MD, MSCE, C. Egla cyclosporine microemulsion to tacrolimus in stable kidney trans- Rabinovich, MD, MPH, Janet Ellis, RT, RDMS, RVT, Janet Wootton, plant patients with hypercholesterolemia is related to an improve- RN, RSCN (Duke University Medical Center, Durham, NC); Peter ment in cardiovascular risk profile: a prospective study. Transplant Chira, MD, MS, Joyce Hsu, MD, MS, Tzielan Lee, MD, Jan Perea, BS, RDMS, RDCS (Stanford University School of Medicine, Palo Alto, 39. Bessant R, Duncan R, Ambler G, Swanton J, Isenberg DA, Gordon CA); Beth Gottlieb, MD, Patricia Irigoyen, MD, Jennifer Luftig, MD, C, et al. Prevalence of conventional and lupus-specific risk factors for Shaz Siddiqi, MD, Zhen Ni, RDMS, RVT, RCDS, Marilynn Orlando, cardiovascular disease in patients with systemic lupus erythematosus: RN, BC, Eileen Pagano, RN, MS (Schneider Children's Hospital, New a case–control study. Arthritis Rheum 2006;55:892–9.
Hyde Park, NY); Andrew Eichenfield, MD, Lisa F. Imundo, MD, 40. Weigel G, Griesmacher A, DeAbreu RA, Wolner E, Mueller MM.
Philip Kahn, MD, Candido Batres, MD, Digna Cabral, BS (Morgan Azathioprine and 6-mercaptopurine alter the nucleotide balance Stanley Children's Hospital of New York-Presbyterian, New York, in endothelial cells. Thromb Res 1999;94:87–94.
NY); Kathleen A. Haines, MD, Suzanne C. Li, MD, PhD, Jennifer 41. Stet EH, De Abreu RA, Bokkerink JP, Blom HJ, Lambooy LH, Weiss, MD, Mary Ellen Riordan, RN, Beena Vaidya, RCDS (Hack- Vogels-Mentink TM, et al. Decrease in S-adenosylmethionine ensack University Medical Center, Hackensack, NJ); Michelle Mietus- synthesis by 6-mercaptopurine and methylmercaptopurine ribonu- Snyder, MD (University of California at San Francisco Medical cleoside in Molt F4 human malignant lymphoblasts. Biochem J Center); Lawrence Ng, BSc (Hospital for Sick Children, Toronto, Ontario, Canada); Susan Ballinger, MD, Thomas Klausmeier, MD, 42. Toloza SM, Uribe AG, McGwin G Jr, Alarcon GS, Fessler BJ, Debra Hinchman, BS, RVT, Andrea Hudgins, CCRP (Indiana Uni- Bastian HM, et al. Systemic lupus erythematosus in a multiethnic versity School of Medicine, Indianapolis); Shirley Henry, RN, PNP, US cohort (LUMINA). XXIII. Baseline predictors of vascular Shuzen Zhang, ARDMS, RVT (Texas Scottish Rite Hospital for events. Arthritis Rheum 2004;50:3947–57.
Children, Dallas); Elizabeth B. Brooks, MD, PhD, Stacy Miner, Nancy 43. Cottiero RA, Madaio MP, Levey AS. Glomerular filtration rate Szabo (University Hospitals/Case Medical Center, Cleveland, OH); and urinary albumin excretion rate in systemic lupus erythemato- Libby Dorfeld, Sarajane Wilson, BS (Children's Hospital of Philadel- sus. Nephron 1995;169:140–6.
phia, Philadelphia, PA); Tatiana Hernandez, BS, Jyotsna Vitale, BS, 44. Bots ML, Evans GW, Riley WA, Grobbee DE. Carotid intima- RT(M), RDMS, RVT (University of California Los Angeles Medical media thickness measurements in intervention studies: design Center); Angela Kress, Nicole Lowe, Falguni Patel (Children's Me- options, progression rates, and sample size considerations: a point morial Hospital, Chicago, IL); Stephanie Hamilton, BSN (Seattle of view. Stroke 2003;34:2985–94.
Children's Hospital and Regional Medical Center, Seattle, WA); Katie 45. Taylor A, Shaw LJ, Fayad Z, O'Leary D, Brown BG, Nissen S, et Caldwell, Diane Kamen, MD, MSCR (Medical University of South al. Tracking atherosclerosis regression: a clinical tool in preventive Carolina, Charleston); Becky Puplava, BS, Atanas Lonchev, RVT, RDCS, cardiology. Atherosclerosis 2005;180:1–10.
RDMS (University of Chicago, Chicago, IL); Monica Bacani, RDMS 46. Espeland MA, Evans GW, Wagenknecht LE, O'Leary DH, Zac- (Nationwide Children's Hospital, Columbus, OH); Cynthia Rutherford, caro DJ, Crouse JR, et al. Site-specific progression of carotid BS, CRCIII, Jamie Meyers-Eaton, BS, CRCIII (Cincinnati Children's artery intimal-medial thickness. Atherosclerosis 2003;171:137–43.
Hospital Medical Center, Cincinnati, OH); Teresa Conway, RN, MS, 47. Schott LL, Wildman RP, Brockwell S, Simkin-Silverman LR, CCRC, Lacey Frank, Lori Kuss, ASRT, RDMS, RDCS (Creighton Kuller LH, Sutton-Tyrrell K. Segment-specific effects of cardio- University Medical Center, Omaha, NE); Hazel Senz, RN, BSN, CCRC vascular risk factors on carotid artery intima-medial thickness in (Denver Children's Hospital, Denver, CO); Thomas Mason, MD, Jane women at midlife. Arterioscler Thromb Vasc Biol 2004;24:1951–6.
Jaquith, CCRC, Diana E. Paepke-Tollefsrud, BS, RVT, RDMS, RT(R) 48. Espeland MA, Tang R, Terry JG, Davis DH, Mercuri M, Crouse (Mayo Clinic, Rochester, MN).

Source: http://oa.smxz.com.cn/upload/uploadfile/200905191242712475.PDF