Cardiovascular risk in pediatric-onset rheumatological diseases

Cardiovascular morbidity and mortality are becoming major health concerns for adults with inflammatory rheumatic diseases. The enhanced atherogenesis in this patient population is promoted by the exposure to traditional risk factors as well as nontraditional cardiovascular insults, such as corticosteroid therapy, chronic inflammation and autoantibodies. Despite definite differences between many adult-onset and pediatric-onset rheumatologic diseases, it is extremely likely that atherosclerosis will become the leading cause of morbidity and mortality in this pediatric patient population. Because cardiovascular events are rare at this young age, surrogate measures of atherosclerosis must be used. The three major noninvasive vascular measures of early atherosclerosis - namely, flow-mediated dilatation, carotid intima-media thickness and pulse wave velocity - can be performed easily on children. Few studies have explored the prevalence of cardiovascular risk factors and even fewer have used the surrogate vascular measures to document signs of early atherosclerosis in children with pediatric-onset rheumatic diseases. The objective of this review is to provide an overview on cardiovascular risk and early atherosclerosis in pediatric-onset systemic lupus erythematosus, juvenile idiopathic arthritis and juvenile dermatomyositis patients, and to review cardiovascular preventive strategies that should be considered in this population.


Introduction
Advances made in the fi eld of pediatric rheumatology over the last decades have led to a signifi cant decrease in mortality rates. Patients are now surviving into adulthood and have to face the many challenges imposed by their chronic illness. As a result of better treatments it is likely that cardiovascular disease will emerge as a leading cause of morbidity and mortality. Th e interplay between traditional cardiovascular risk factors, exposure to corticosteroids and chronic infl ammation creates a perfect storm for early atherogenesis in this population.
Atherosclerosis is now being increasingly recognized in adults with infl ammatory rheumatic diseases but there is little information pertaining to pediatric-onset rheumato logic conditions [1]. Th e objectives of this review are to summarize the current state of knowledge on cardiovascular risk and accelerated atherosclerosis in pediatriconset systemic lupus erythematosus (pSLE), juvenile idiopathic arthritis (JIA) and juvenile dermatomyositis (JDM), and to discuss atherosclerosis preventive strategies that should be considered in this patient population.

Measuring atherosclerosis in children
Although cardiovascular events are the true markers of atherosclerosis in rheumatic diseases, including those in adults, examining cardiovascular risk factors and/or preventive strategies for atherosclerosis are limited by sample size, and therefore surrogate outcome markers (vascular measures of early atherosclerosis) are required. Currently there are three major vascular markers that have been validated as measures of early atherosclerosis: fl ow-mediated dilatation (FMD), carotid intima-media thickness (CIMT), and pulse wave velocity (PWV).
Endothelial injury is an important initial event in the development of atherosclerosis and therefore measurement of endothelial function can serve as a surrogate marker of atherosclerosis [2]. Vascular ultrasound of the brachial artery under conditions of FMD after reactive hyperemia (endothelium-dependent vasodilatation) and in response to glyceryl trinitrate (endothelium-independent vasodilatation) are noninvasive techniques to measure endothelial function. Th e severity and longterm risk of coronary artery disease has been well correlated with FMD [3,4].
Ultrasound study of CIMT is a reproducible, validated measurement, and increased CIMT and the presence of Abstract Cardiovascular morbidity and mortality are becoming major health concerns for adults with infl ammatory rheumatic diseases. The enhanced atherogenesis in this patient population is promoted by the exposure to traditional risk factors as well as nontraditional cardiovascular insults, such as corticosteroid therapy, chronic infl ammation and autoantibodies. Despite defi nite diff erences between many adult-onset and pediatric-onset rheumatologic diseases, it is extremely likely that atherosclerosis will become the leading cause of morbidity and mortality in this pediatric patient population. Because cardiovascular events are rare at this young age, surrogate measures of atherosclerosis must be used. The three major noninvasive vascular measures of early atherosclerosis -namely, fl ow-mediated dilatation, carotid intima-media thickness and pulse wave velocity -can be performed easily on children. Few studies have explored the prevalence of cardiovascular risk factors and even fewer have used the surrogate vascular measures to document signs of early atherosclerosis in children with pediatric-onset rheumatic diseases. The objective of this review is to provide an overview on cardiovascular risk and early atherosclerosis in pediatric-onset systemic lupus erythematosus, juvenile idiopathic arthritis and juvenile dermatomyositis patients, and to review cardiovascular preventive strategies that should be considered in this population.
plaques are predictive of future coronary artery disease and stroke [5]. Meta-analyses and reviews have repeatedly shown that CIMT can predict the risk of future cardiac events and that change over time and a reduction in CIMT is associated with a reduction in incidence of cardiovascular disease (CVD) events [6,7].
PWV is a noninvasive, reliable and reproducible way of measuring early changes in arterial wall stiff ness and arterial distensibility [8]. Increased PWV has been demon strated in both coronary artery disease and cerebrovascular disease [9].
Although data are still lacking on the predictive value of these surrogate markers for future events in pediatric patients, a scientifi c statement from the American Heart Association stated that these studies detect subclinical vascular disease and therefore can identify children at risk for CVD [10]. Endothelial dysfunction may be found in multiple pediatric rheumatologic diseases and its eff ects on vascular markers need to be better characterized. Abnormal vascular measures could result from transient, infl ammation-induced endothelial dysfunction and not from atherosclerosis per se. Interpretation of these surrogate markers should be made with caution until better methods of distinguishing these phenomena are developed.

Pediatric-onset systemic lupus erythematosus
PSLE is a life-long autoimmune disease characterized by chronic infl ammation, the production of autoantibodies and the frequent use of corticosteroid therapy. Disease onset before age 18 (pSLE) accounts for approximately 15% of all cases of systemic lupus erythe matosus (SLE) [11]. Cardio vascular disease is a leading cause of morbidity and mortality in adult SLE [12]. Although there has been a signifi cant improvement in standardized all-cause mortality rates in adult SLE patients over time, the mortality secondary to athero sclerosis and cardiovascular diseases has not signifi cantly changed [13].
One of the few studies to report on the long-term burden of CVD in pSLE, the Lupus Outcome Study demonstrated that patients with onset of SLE during childhood not only had a similar incidence of myocardial infarction (MI) to subjects with adult-onset disease but that the fi rst MI occurred at a much earlier mean age (32 years) [14]. Using data from this study and agematched cardiovascular data, by early adulthood pSLE patients have a 100-fold to 300-fold increased risk of death from CVD as compared with age-matched controls [15,16]. However, large prospective long-term cohort studies following patients with pSLE into adulthood are required to determine the true extent of cardiovascular morbidity and mortality. Until these collaborative studies are performed, we need to rely on markers of vascular function as surrogates for atherosclerotic events.

Evidence of early atherosclerosis in pSLE
Because many years of exposure to atherosclerosis risk factors are required before a cardiovascular event occurs, we cannot rely on the incidence of hard outcomes to assess for atherosclerosis in pSLE and other pediatric rheumatic diseases. Th e noninvasive vascular assessment techniques of CIMT, FMD and PWV described above therefore play a central role in the detection of early atherosclerosis in pSLE. Th e major problem when compar ing and reviewing studies in pSLE is the heterogeneity of the population studied, the small sample sizes of each study, the diff erent covariates included in the analysis and the diff ering vascular study protocols used (Table 1) [17][18][19][20][21][22][23]. Th ese factors may lead to both type I and type II errors when determining the incidence and prevalence of early atherosclerosis in pSLE. However, despite these limitations, certain conclusions can be made. As seen in studies of adult SLE, PWV may be the best measure to detect the earliest changes of atherosclerosis in pSLE. Furthermore, it has been suggested that PWV and CIMT may measure diff erent vascular damage [24].
Cross-sectional studies of CIMT reported varied results as compared with age-matched controls, probably as a result of the diff ering vascular study protocols used. Despite these limitations, the average annual rate of progression of CIMT in pSLE patients is small (between 0.0016 and 0.0024 mm/year) ( Table 2). Th is rate of change in CIMT is similar to that in a study of 247 healthy Caucasian adolescents, which found a rate of change of 0.02 mm over a 10-year period (average 0.002 mm/year) [25]. In contrast, in the Atherosclerosis Prevention in Pediatric Lupus Erythematosus (APPLE) trial, the progression rates of all but one of the diverse CIMT measurements carried out on the enrolled placebo-treated pSLE patients showed higher numerical values than those reported in the healthy adolescent cohort, suggesting accelerated athero sclerosis in this large pSLE cohort [26]. More large-scale longitudinal studies are required to determine whether CIMT progression in pSLE truly follows an abnormal trajectory. Adjustment for factors that infl uence CIMT in childhood, such as age, height, body mass index and blood pressure, should be made to ensure appropriate conclusions are drawn [25]. Elevated PWV has been described in pSLE patients [17]. Interestingly, this later study was carried out on pSLE subjects relatively early in their disease course who had low disease activity. Despite this, a signifi cant diff erence from healthy controls was found. Th is might suggest that atherogenesis starts early in these patients and may aff ect even those with low disease activity.
Measures of brachial artery reactivity (FMD) show the most variation and the signifi cance of these measurements is not clear. Most studies have examined patients with short disease duration and therefore they may not refl ect the long-term burden of atherosclerosis in pSLE subjects [17,23].

Risk factors for atherosclerosis
Even though factors associated with progression of subclinical atherosclerosis have not been well explored in pSLE, traditional and nontraditional risk factors are being increasingly recognized to be present in this patient population.

Traditional risk factors for cardiovascular disease in pSLE
Dyslipidemia, hypertension, obesity, sedentary lifestyle, diabetes, smoking and family history of early CVD are all con sidered traditional risk factors for atherosclerosis develop ment. As compared with adult SLE patients, uncontrolled hypertension, diabetes and smoking are not commonly encountered in pediatric patients. In contrast, dyslipidemia -one of the key precursors of early atherosclerosis -has been shown to be present with increased frequency in pSLE when compared with the general pediatric population.

Dyslipidemia
As early as 1988 it was recognized that both adult and pediatric SLE patients with active disease, prior to corticosteroid therapy, had elevated triglycerides (TGs) and very-low-density lipoprotein-cholesterol and had depressed high-density lipoprotein-cholesterol (HDL-C) and apolipoprotein A-I -often referred to as the active lupus lipid profi le [27,28]. Th e lipid profi le of elevated TGs and depressed HDL-C is not specifi c for SLE but rather is common to multiple infl ammatory states.
Later studies showed that lipid abnormalities were common in newly diagnosed pSLE patients, prior to the start of corticosteroid treatment [29]. Among a group of 54 recently diagnosed and untreated pSLE subjects, at least one lipid abnormality was found in 63% of patientselevated TGs in 62%, abnormally low HDL-C in 24%, hypercholesterolemia in 20% and elevated low density lipoprotein cholesterol (LDL-C) in 4% -and one patient had an abnormal level of all four lipids [29].
Th e cross-sectional APPLE trial found mean TG, LDL-C and HDL-C levels that were in the normal ranges [30]. Th e most common lipid abnormality was also elevated TG levels, found in nearly 30% of subjects.
Th e diff erences between these two studies are probably the result of the lower disease activity, longer disease duration and immunosuppressive therapies of patients in the APPLE trial as compared with the untreated, active patients in the former study.
Th e dyslipidemia seen in pSLE patients is multifactorial. Cytokines, autoantibodies, medications, dietary intake, renal disease, physical activities and genetics factors are all likely important contributors. As in other diseases, nephrotic-range proteinuria is associated with higher Table 1

Falaschi and
Bowser and Huang and Schanberg and Boros and colleagues [19] colleagues [18] colleagues [20] colleagues [22] colleagues [  total cholesterol and LDL-C levels in pSLE [19]. In addition, other investigators have shown that lower levels of proteinuria (defi ned as ≥200 mg/day but <3.5 g/day) were associated with abnormal lipid levels, again demonstrating the importance of renal disease in the dys lipidemia of pSLE [31].
Th ere are few longitudinal studies that have examined serial lipid levels in pSLE subjects [31,32]. In an inception cohort of 139 pSLE patients followed annually from diagnosis up to 3 years, it was shown that the highest levels of total cholesterol, LDL-C and TG and the lowest levels of HDL-C were found at diagnosis, prior to onset of therapy [32]. On follow-up, these same investigators found that the most important factors infl uencing lipid levels were changes in disease activity and prednisone dose [31]. When controlling for disease activity, a reduction in the prednisone dose was associated with an overall improved lipid profi le. A key concept emerges from these studies: when disease activity is brought under control, steroid therapy is being weaned and proteinuria improves, the lipid values normalize. Th ese results suggest that disease control rather than long-term lipid-lowering therapy may be the most important factor to control dyslipidemia in pSLE.

Insulin resistance and metabolic syndrome
Both insulin resistance and metabolic syndrome are present in SLE [32][33][34][35]. In pSLE, fasting insulin levels, hemoglobin A 1 C and insulin C-peptide levels were elevated, although Homeostatic Model Assessment for Insulin Resistance values were normal in the majority of patients [17,36]. Obesity, chronic infl ammation, and corticosteroid therapy all can contribute to insulin resistance [37][38][39]. However, elevated fasting insulin levels and hyperinsulinemia were not restricted to patients on corticosteroids or to obese patients. Insulin resistance may also be secondary to a chronic infl ammatory state including SLE. A rare cause of insulin resistance (type B insulin resistance) is the presence of antibodies against the insulin receptor. Th ese antibodies result in hyper gly cemia, insulin resistance and acanthosis nigricans, and the majority of patients with anti-insulin receptor antibodies have SLE [40]. Of note, patients with anti-insulin receptor antibodies may present with and/or develop hypoglycemia. Despite the presence of insulin resistance and/or metabolic syndrome in patients with pSLE, there is no evidence of an increased prevalence of overt type I or type II diabetes mellitus in pSLE or adult SLE [41]. Patients are, however, at risk for steroid-induced diabetes.

Nontraditional risk factors for cardiovascular disease in pSLE
In adult SLE patients, traditional risk factors alone are not suffi cient to explain the enhanced risk of cardiovascular events. Indeed, after controlling for the Framingham risk factors, SLE patients have a 17-fold increased risk of mortality from coronary heart disease and a 10-fold increase risk of nonfatal MI [42].

Hyperhomocysteinemia
In studies of otherwise healthy individuals it is not clear whether an elevated homocysteine plasma level is a causal factor or simply a biomarker for atherosclerosis. An etiologic role of hyperhomocysteinemia in atherosclerosis is supported by the multiple proatherogenic eff ects of hyperhomocysteinemia: induction of endo thelial cell dysfunction; increased oxidative stress; and promotion of transcription of proinfl ammatory cytokine genes via activation of the NF-κB pathway [43][44][45]. However, the failure of most trials of homocysteinelowering therapy to decrease cardiovascular risk in large randomized studies has cast doubt on the role of hyperhomocysteinemia in atherosclerosis [46,47]. Studies of both pSLE and adult SLE patients have shown elevated plasma homocysteine levels. However, there was no correlation between elevated homocysteine levels and vascular markers of early atherosclerosis [17,48,49].

Cytokines and adipokines
An imbalance between endothelial cell damage and repair has been demonstrated to occur in SLE [50,51]. Th ese abnormalities as well as recruitment of macrophages to arteries, enhanced formation of foam cells and platelet activation have all been shown to be induced by type I interferons [52][53][54]. Increased serum type I interferon activity was shown to be a predictor of abnormal FMD and increased CIMT in adult SLE patients [55]. Similar studies have not been conducted in pSLE but, as the interferon signature is seen in pSLE [56], type I interferon may be an important disease-related factor promoting atherosclerosis.
Elevated levels of other multiple proinfl ammatory cytokines, including IL-6, IFNγ and TNFα, have been implicated in the development of atherosclerosis in other wise healthy populations [57][58][59]. Although elevated levels of IL-6 are present in adults with SLE and these levels have been shown to correlate with C-reactive protein and abnormal lipid levels, no association between IL-6 and the presence or progression of atherosclerosis has been found in SLE [60][61][62][63][64].
Adiponectin, leptin and ghrelin all belong to the adipokine family of molecules. Synthesized by adipous tissue, these molecules infl uence energy homeostasis, lipid metabolism and insulin sensitivity and regulate appetite. Evidence is emerging that adiponectin might be a protective mediator of atherosclerosis, because in non-SLE populations lower levels of adiponectin were associated with a higher risk of MI [65]. In contrast, high levels of adiponectin were observed in adult SLE and other infl ammatory states. However, there has been no consistent correlation between adipokine levels and early markers of atherosclerosis in adult SLE [66][67][68][69]. A prospective study of pSLE patients found no signifi cant diff erence in adiponectin levels between 105 pSLE patients and a group of healthy controls, although seven pSLE subjects had elevated levels [70]. No studies have explored factors associated with changes in adipokine levels over time or the predictive value of these molecules for the progression of atherosclerosis in pSLE.
Further studies are needed to defi ne whether cytokines, chemokines and adipokines drive the development and progression of atherosclerosis or are protective. Th ese potential biomarkers could guide physicians in stratifying CVD risk in pSLE patients.

Autoantibodies
Lupus is characterized by circulating antibodies of multiple specifi cities, including anti-phospholipid, antiendothelial cells, anti-apolipoprotein A-I, anti-HDL-C, anti-lipoprotein(a), anti-oxidized LDL-C and anti-lipopro tein lipase antibodies. Th e presence of these autoantibodies against these key constituents in the atherosclerosis cascade has been postulated to contribute to the pathogenesis of early atherosclerosis in SLE. In pSLE patients, both antiphos pholipid and anti-oxidized LDL-C antibodies have been found [23,71].

Chronic kidney disease
Lupus nephritis is a common disease manifestation of pSLE. In a minority of patients, renal involvement leads to altered glomerular fi ltration rate and even to end-stage renal disease. Although not a traditional cardiovascular risk factor, chronic kidney disease is a major contributor to the onset and progression of accelerated atherosclerosis in this subgroup of individuals. Analysis of the US Renal Data System, a database comprising patients on chronic dialysis therapy or patients who had received a renal transplantation, revealed that the risk of death was twice as high in pSLE end-stage renal disease patients when compared with pediatric patients with non-SLErelated end-stage renal disease [72]. Seventy-fi ve percent of those deaths were due to cardiovascular disease and cardiac arrest.
A correlation between the duration of end-stage renal disease and CIMT as well as improvement of CIMT post renal transplantation has been described in pediatric chronic kidney disease patients [73,74]. Similar risk factors for atherosclerosis may be found in both SLE and chronic kidney disease patients: hypertension, dyslipidemia, a chronic infl ammatory state, oxidative stress and underlying endothelial dysfunction [75][76][77][78][79]. Th e coexistence of this double hit and of other chronic kidney disease-specifi c risk factors, including activation of the renin-angiotensin-aldosterone and sympathetic nervous systems as well as enhanced vascular calcifi cation due to metabolic derangements, lead to a strong proatherogenic milieu. Special attention should therefore be given to pSLE patients with chronic kidney disease as they are at increased risk of cardiovascular disease.
Th ere is a signifi cant gap in knowledge on factors leading to accelerated atherosclerosis and long-term cardio vascular outcomes in pSLE. Identifying key predictors of CVD in this vulnerable population is crucial. Th is will enable pediatric rheumatologists to identify early on patients at highest risk and to prioritize the imple mentation of preventive strategies.

Juvenile idiopathic arthritis
JIA is the most common rheumatologic disease of childhood. JIA encompasses diverse disease subtypes with marked variation in the level of systemic infl ammation. Adults with multiple forms of infl ammatory arthritis have been shown to have premature atherosclerosis and an increased morbidity and mortality secon dary to atherosclerosis [80]. In fact, it has been shown that patients with rheumatoid arthritis probably have a similar risk for atherosclerosis as those with type 2 diabetes mellitus [81]. Despite the diversity of the clinical manifestations of the diff erent JIA subsets, most patients have evidence of chronic infl ammation and therefore are probably at risk for early atherosclerosis. Diff erent disease subtypes probably have distinct risk profi les.
Th ere is a paucity of data on the risk or outcome of artherosclerosis in JIA patients with respect to morbidity and mortality. One autopsy study had suggested that 30% of children with JIA had evidence of atherosclerosis [82]. A small study from Poland showed that the mean CIMT was within the published normal range although increased CIMT was associated with disease duration [83,84]. A review of the English-language literature revealed only three reports that assessed surrogate measures of atherosclerosis in children with JIA (Table 3) [85][86][87]. Th e fi rst reported study, using magnetic resonance imaging, showed that JIA patients had increased aortic stiff ness, lower aortic distensibility and increased PWV, as compared with controls, suggestive of subclinical atherosclerosis [85]. Th e only association with increased PWV was age. A subsequent comprehensive study showed normal CIMT and PWV but abnormal FMD in JIA patients. Post-hoc analysis suggested that the abnormal FMD was probably related to the infl ammatory state of the JIA patients at time of measurement. Th ere was no diff erence between patients with active and inactive JIA for any of the vascular markers. Not surprisingly, patients with systemic JIA -the subtype of JIA characterized by the most signifi cant chronic infl am mation -had the highest CIMT. However, only seven systemic JIA patients were tested [87]. Th e most recent study reported that prepubertal JIA patients (only oligoarticular and polyarticular subtypes) had significantly increased CIMT as compared with controls. At follow-up after 1 year of individualized therapy that was associated with improved disease control and decreased infl ammation, there was a signifi cant decrease in CIMT [86]. In this study, CIMT was associated with LDL-C and IL-1 levels. Overall, the small number of patients and the predominance of subjects with oligo articular JIA, the least infl ammatory subtype, make the generalizability of these fi ndings questionable. Children with systemic JIA are probably at a much higher risk of accelerated atherosclerosis than those with oligoarticular or polyarticular JIA. Future studies should address the question by specifi c disease subtypes.
Similar issues arise when studies on the prevalence of traditional risk factors for atherosclerosis are examined. Th e few studies reporting lipid levels of JIA patients included children with diff erent disease subtypes, disease activity levels and corticosteroid doses. Th is heterogeneity probably explains the contradictory fi ndings of the active infl ammatory lipid profi le of high TGs and very-low-density lipoprotein-cholesterol and lower levels of HDL-C, LDL-C and total cholesterol found in some, but not all, studies of JIA [86][87][88][89][90][91][92][93]. Overall it is diffi cult to determine the eff ect of disease and disease activity on the lipid profi le in each distinct JIA subtype. Moreover, biologic therapies might adversely impact the lipid profi le, as seen in systemic-onset JIA patients treated with tocilizumab [94]. Further large-scale studies should be conducted to clarify the burden and trajectory of dyslipidemia in the diff erent JIA subtypes. Similar to other rheumatic diseases, patients with JIA have evidence of elevated proinfl ammatory cytokines and chemokines. Th e degree of elevation is dependent on the subtype and level of disease activity. Most, but not all, studies have shown elevated homocysteinemia levels in JIA patients [95][96][97]. One study showed a correlation of homocysteine with CIMT [83]. Elevated levels of omentin, an insulin-sensitizing adipokine, were demonstrated in patients with JIA [98]. Omentin has important eff ects on endothelium as it causes vasodilatation of blood vessels and attenuates many proinfl ammatory signal ing pathways in endothelial cells, including the TNF pathway [99,100]. Th e elevated circulating levels seen in JIA may therefore act as a counter-regulatory mecha nism to attenuate the proatherosclerotic eff ects of TNF and other mediators of chronic infl ammation. Th e importance of leptin in atherosclerosis in JIA is not clear because the two studies reported showed opposite results [101,102].
Limitation to performing physical activities due to arthritis and its associated musculoskeletal complications, as well as corticosteroid usage, renders JIA patients susceptible to weight gain. A recent study found that patients with JIA had higher body mass index, fat percentages and truncal fat than age-matched and sexmatched controls [103]. Despite the challenges to regular physical activities in this patient population, children with JIA should be strongly encouraged to engage in an exercise program.
Data on other nontraditional risk factors in this pediatric population are lacking. As many patients will continue to have active disease into adulthood, the need to expand research on the prevalence of atherosclerosis Values are presented as mean ± standard deviation or median (range), unless otherwise specifi ed. CIMT, carotid intima-media thickness; FMD, fl ow-mediated dilatation; JIA, juvenile idiopathic arthritis; PWV, pulse wave velocity. a Diff erence from controls is statistically signifi cant (P <0.01). b Diff erence from controls is statistically signifi cant (P = 0.001).
risk factors and incidence of CVD in JIA is important. Th is is especially true for patients with systemic and polyarticular JIA subtypes, as these patients are at highest risk for chronic infl ammation and prolonged use of corticortisteroids.

Juvenile dermatomyositis
JDM is an infl ammatory disease characterized by typical cutaneous rashes, symmetrical proximal muscle weakness, raised serum levels of muscle enzymes and a vasculo pathy. Few studies have assessed early atherosclerosis in JDM patients. A study of adults with dermatomyositis or polymyositis demonstrated higher risks of acute MI and stroke when compared with the general population [104]. A recent study from Sweden showed that adults with polymyositis/dermatomyositis had a standardized incidence ratio of 1.92 for the development of coronary heart disease as compared with age-and sexmatched controls [105]. Th e only study to examine the risk of atherosclerosis in JDM compared CIMT and FMD of eight adults with a history of JDM with eight healthy adults [106]. Th e CIMT was higher in JDM patients despite the fact that they were younger and had lower body mass index than controls. Acquired lipodystrophy has gained recognition in JDM patients with reported prevalence rates ranging from 8 to 40% [107][108][109]. Th is condition leads to loss of sub cutaneous fat and is associated with diverse metabolic abnormalities such as dyslipidemia, abnormal leptin levels, insulin resistance and overt diabetes. Th e extent to which these abnormalities impact on accelerated atherosclerosis remains unknown. Future work should evaluate whether lipodystrophy is a predictor of cardiovascular events in the JDM population.
A study of children with myositis showed that 41% had elevated fasting insulin levels with 25% meeting criteria for metabolic syndrome, 47% had elevated TG and 17% had abnormal cholesterol, LDL-C or HDL-C levels [110]. Th e previous mentioned study of CIMT was also performed to document risk factors for atherosclerosis in JDM [106]. Th ese same patients had higher blood pressure values, higher prevalence of abnormal HDL-C levels and lower adiponectin levels compared with controls. Two patients fulfi lled criteria for the metabolic syndrome and 63% had lipodystrophy [107]. Atherogenic risk factors are thus clearly present in JDM children. Again, future studies should address the long-term incidence of CVD and better delineate predictors of cardio vascular morbidity and mortality in JDM.
Although JDM and pSLE are clearly two distinct entities, a parallel can be established between these conditions with respect to cardiovascular risk factors. Type I interferons have been implicated in the pathogenesis of both JDM and SLE [111]. Although no studies have looked at the eff ect of these cytokines on the development of atherosclerosis in JDM, it is reasonable to assume that they have similar eff ects on the endo thelium to those seen in SLE. Similarly, chronic insult to the endothelium precipitated by the underlying vasculo pathy and chronic infl ammation probably leads to early atherosclerosis in JDM, as seen in SLE. Additionally, long-term corticosteroid therapy probably results in similar metabolic derangements such as weight gain, insulin resistance and dyslipidemia. Both diseases also lead to disabilities resulting in a decreased ability to practice regular aerobic activities.

Prevention strategies
Atherosclerosis is known to begin in childhood. Moreover, longitudinal studies have demonstrated that the presence of cardio vascular risk factors in childhood is associated with higher CIMT in young adults [112,113]. Children with infl ammatory diseases have an increased prevalence of traditional risk factors and the extra burden of non traditional risk factors for atherosclerosis. Hence, thorough ongoing cardiovascular risk assessment should be performed routinely in children with chronic rheumatologic diseases. Long-term cardiovascular morbidity and mortality should be discussed with patients and their families soon after disclosure of the diagnosis to educate patients and facilitate the application of preventive strategies.

Traditional risk factor-related preventive measures
Overweight and obesity are common in patients with pediatric rheumatic diseases, and in particular in patients on corticosteroid therapy. Th e general unwellness secondary to systemic infl ammation, the arthritis, the psychological impact of having a chronic disease and the development of cushingoid features all potentially contribute to sedentary lifestyle. Th e benefi ts of physical activity on many of the traditional cardiovascular risk factors are well established and regular exercise should be encour aged. Children unable to perform aerobic exercises to the desired level should be encouraged to pursue a regular exercise routine adapted to their clinical status, and intensity should be increased as tolerated. A healthy balanced diet should be promoted in all patients. Nutritional and physical activity counseling should be regarded as management priorities. Smoking status should be assessed regularly and smoking cessation support off ered.
Blood pressure should be monitored at each clinic visit and home blood pressure monitoring should be implemented in those with borderline elevated or elevated values or with kidney disease. Periodic monitoring of the blood glucose level should be performed with special attention to patients on a high dosage of cortico steroids, overweight or obese, or with a family history of type II diabetes mellitus.
A baseline pretreatment lipid level should be measured and serial measurements performed. Th is may not apply to certain disease subtypes, such as oligoarticular JIA, where the risk of dyslipidemia is probably similar to that in the general pediatric population. However, there are currently no clear guidelines on the frequency of dyslipidemia screen ing and the threshold at which specifi c treatment should be contemplated for patients with pediatric rheumatic diseases has not been defi ned [114]. Th e APPLE trial, the only study to pros pectively assess the use of a statin to reduce the progression of atherosclerosis, as measured by CIMT in pSLE, did not fi nd a signifi cant diff erence in progression of CIMT between the statin-treated and placebo-treated patient groups [26]. However, secondary analyses did show a trend in favor of atorvastatin for other CIMT endpoints. Further studies are required to determine the role of lipid-lowering therapy in pediatric rheumatology.
Th e use of antimalarial agents in pSLE and JDM provides a dual benefi t: not only do these agents help to keep the disease inactive, but they also exert benefi cial eff ects on lipid levels and glucose tolerance [115][116][117][118].

Nontraditional risk factor-related preventive measures
A key message that emanates from SLE studies and that probably applies to other infl ammatory conditions is that adequately treating the primary disease will improve many of the cardiovascular risk factors. Control of sys temic infl ammation will decrease production of pro athero genic cytokines, chemokines, adipokines and auto anti bodies, thereby decreasing the burden of pro athero genic insults. Judicious use of corticosteroid might tilt the balance in favor of benefi ts. Th is hypothesis might explain why the use of moderate doses of corticosteroid was negatively correlated with CIMT in the APPLE study [26].
Th erapies used in the treatment of rheumatic disease have been associated with a decrease in atherosclerotic burden. Anti-TNFα therapies have been associated with decreased CIMT [119]. Although B cells have been found to have atheroprotective eff ects, B-cell defi ciency in murine models of atherosclerosis results in a decrease in plaque size [120,121]. Whether this eff ect is also true in humans and whether other anti-B-cell treatments have the same eff ect are not clear. As there are increasing data on the importance of type I interferons in atherosclerosis and disease activity, it will be interestingly to see whether anti-interferon therapy will aff ect atherosclerosis. Th ese biologic thera peutic avenues off er the potential for additional tools in atherosclerosis prevention and therapy. Whether the primary eff ect on atherosclerosis is secondary to decreas ing infl ammation or whether there are intrinsic factors related to these molecules is not clear.
Azathioprine has been linked to atherosclerosis while methotrexate and mycophenolate mofetil may off er protec tion [122][123][124][125]. However, the eff ects of these three immuno suppressive agents on atherosclerosis need further study. Th e use of angiotensin-converting enzyme inhibitors and angiotensin receptor blocking agents in patients with proteinuria and/or hyperten sion will not only off er benefi ts by their direct action on these cardiovascular risk factors but probably also via downregulation of the renin-angiotensin system, which is also implicated in the genesis of atherosclerosis.

Conclusion
Children with chronic rheumatologic diseases are exposed to a vast array of proatherogenic insults, but the prevalence and natural history of accelerated atherosclerosis in the majority of these children remains poorly defi ned. However, it is becoming more apparent that cardiovascular disease results in signifi cant morbidity and mortality in these patients in adulthood. Identifying key risk factors, developing disease-specifi c stratifi cation algorithms and implementing interventions to prevent atherosclerosis are therefore important. Th e predictive value of surrogate measures of athero sclerosis should be specifi cally studied in this pediatric population. Eff orts should be made to identify novel biomarkers that would assist us in quantifying the atherosclerotic burden and to follow its trajectory. Prospec tive, multicenter cohort studies addressing these impor tant issues are urgently needed. Pediatric rheumatology researchers have shown that assembling large national and international cohorts of patients with rheumatic diseases is feasible [15,126,127]. Hopefully, collaborative eff orts among the pediatric and adult rheumatology communities will ultimately lead to improved long-term cardiovascular outcomes in patients with pediatric-onset rheumatologic diseases.