Introduction
Physicians who care for children with chronic kidney disease (CKD) increasingly recognize that cardiovascular disease is a major threat to their patients. Several reports have confirmed that, as in adults, cardiovascular events are the main cause of death among patients who develop end-stage renal disease (ESRD) in childhood.1, 2, 3, 4 In a statement by the American Heart Association on cardiovascular risk reduction in high-risk pediatric patients, those with CKD were classified in the highest risk stratum, alongside individuals with homozygous familial hypercholesterolemia, diabetes mellitus type 1, heart transplants or coronary aneurysms due to Kawasaki disease.5
Renal failure is associated with a cluster of classic risk factors for cardiovascular disease. In addition, uremia-specific risk factors seem to contribute to cardiovascular risk in this setting.6 In adults with renal failure, accelerated atherosclerosis leads to myocardial infarction and contributes substantially to cardiovascular mortality. In pediatric patients, risk factors for accelerated atherosclerosis are present, but cardiovascular mortality is largely the result of ESRD-specific factors.
In this Review, we discuss the arterial and cardiac abnormalities found in children with CKD and ESRD, the potential determinants of these abnormalities, and their prevention.
Mortality and morbidity
Complete and reliable data on the late outcomes of children with ESRD are scarce (Table 1). Two studies on mortality and causes of death in this setting reported long-term follow-up data from patients who started renal replacement therapy (RRT) during childhood, with a mean follow-up of 9.7 and 16.0 years, respectively.1, 2 In addition, Oh et al.7 studied mortality and causes of death among all 283 patients treated for ESRD at a single pediatric center. Nearly 6% of this cohort was lost to follow-up. Parekh et al.4 studied the causes of death of 1,380 patients included in the United States Renal Data System who started RRT before the age of 20 years and died between 1990 and 1996, before the age of 30 years. Chavers et al.3 analyzed the causes of 107 deaths in Medicare patients who started dialysis between 1991 and 1996, while under the age of 20 years.
Table 1 | Cardiovascular outcomes in children with chronic kidney disease or end-stage renal disease
Overall mortality in children with ESRD is about 30 times higher than that among children without ESRD,1, 2 and the risk of cardiac death is more than 700-fold greater.8 When hyperkalemia is excluded as a cause of mortality, cardiovascular disease1, 2, 4, 7 and cardiac death1, 2, 4 account for about 40% and 20–30% of all deaths, respectively.1, 2, 4, 7 Uncertainty about the exact cause of cardiac death in children with ESRD is reflected by the high percentage of diagnoses of 'cardiac arrest' (13–52% of all cardiac deaths, excluding cerebrovascular death and hyperkalemia1, 2, 4). This apparently high incidence of cardiac arrest might be caused by inaccurate labeling of myocardial infarction and/or arrhythmia due to myocardial hypertrophy. Cerebrovascular death occurred in children on hemodialysis at higher rates in the early era of RRT than it does now.1, 2 Although overall mortality decreased during the first few decades of RRT, it has stabilized since the early 1990s.1 Risk factors for death in children with ESRD include dialysis (versus transplantation),1, 2, 4 sustained hypertension2 and young age of onset of RRT.1, 2, 4 Black ethnicity is associated with an increased risk of cardiac death.3, 4
Chavers et al.3 also reported on 452 cardiovascular events in 1,454 children (31%) with ESRD. The overall incidence of these events increased from 24.3% in the 0–4-year age-group to 36.9% in the 15–19-year age-group. Arrhythmia was the most common cardiac event (19.6%), followed by valvular heart disease (11.7%), cardiomyopathy (9.6%) and acute cardiac death (2.8%). From 1991 to 1996, the rate of cardiomyopathy increased significantly (from 4.2 to 8.5 cases per 100 patient-years, P = 0.003), but the rates of other cardiac events remained unchanged. The design of this study precluded the authors from providing an explanation for this phenomenon.3
Pathophysiology and risk factors
Endothelial dysfunction, accelerated atherosclerosis, arterial stiffening due to vascular calcification, left ventricular hypertrophy (LVH) and myocardial fibrosis can occur simultaneously in adults with advanced renal disease. All these abnormalities have their own determinants, and they also strongly influence each other's development. Endothelial dysfunction exacerbates arterial luminal narrowing and arterial wall stiffening by allowing the development of intima–media thickening, medial hypertrophy and calcification. These processes increase the effect of specific renal-disease-associated risk factors for vascular calcification. Arterial wall calcification leads to a loss of wall distensibility, which increases cardiac pressure load, and hence induces LVH. Ultimately, LVH and myocardial fibrosis can provoke sudden cardiac death by arrhythmia or cardiac failure. Calcification of the cardiac valves might further contribute to LVH. The combination of LVH and accelerated atherosclerosis puts adults with ESRD at high risk of early myocardial infarction. In this section, we discuss the roles and potential determinants of the above-mentioned mechanisms in pediatric renal failure.
Endothelial dysfunction
The endothelium is a key factor in the maintenance of vascular tone and in prevention of thrombosis and local inflammation of the vascular wall. Endothelial dysfunction is considered to be the initiating step in the genesis of atherosclerosis and arteriosclerosis,9 and such dysfunction predicts cardiovascular morbidity and mortality.10 In adults, endothelial dysfunction appears early in renal disease and has been attributed to a number of potential causes: reduced clearance of the endothelial nitric oxide synthase (eNOS) inhibitor asymmetric dimethylarginine (ADMA), which leads to reduced bioavailability of endothelial nitric oxide; activation of angiotensin II, which induces oxidative stress; high levels of homocysteine; chronic inflammation; and dyslipidemia.11, 12 Impairment of endothelial repair mechanisms might also contribute to endothelial dysfunction. Repair of damaged endothelium seems to depend not only on resident cells in the vascular wall, but also on bone-marrow-derived circulating endothelial progenitor cells (EPCs). Low circulating numbers and disturbed in vitro function of EPCs have been found in adults with cardiovascular disease and in patients with ESRD.13 Moreover, the number of circulating EPCs predicts cardiovascular mortality in adults in the general population.14 Little is known about circulating EPCs in children, although healthy children have higher numbers of circulating EPCs than healthy adults, possibly as a result of active growth.15 Whether renal failure influences EPC number and function in children remains to be determined.
As in adults, endothelial dysfunction appears early in renal disease in children, and has been observed in children with conservatively treated CKD as well as in children on dialysis and in those with a kidney transplant.16, 17, 18, 19, 20, 21, 22 Also as in adults, the accumulation of endogenous inhibitors of eNOS seems to be responsible in part for endothelial dysfunction.18, 20 Data on the influence of dialysis on endothelial function in adults are conflicting, but in children, Lilien et al.19 demonstrated that flow-mediated vasodilatation deteriorated acutely after a hemodialysis session. Resistance to growth hormone might contribute to endothelial dysfunction in children, as flow-mediated dilatation also decreased after temporary cessation of growth hormone therapy in children with CKD.23
Atherosclerosis and vascular calcification
In adults with ESRD, calcifications are found in both the intimal and medial layers of the arterial wall. Intimal calcification is the result of advanced atherosclerosis, which is characterized by the formation of irregular, calcified plaques and occlusive lesions. Intimal calcification is most commonly found in older patients with ESRD. By contrast, diffuse, nonocclusive, medial calcification (Monckeberg's arteriosclerosis) is more dominant than intimal calcification in young adults with ESRD.24 Medial calcification increases arterial stiffness, and thus reduces the elasticity and compliance of arteries. This loss of compliance increases systolic pressure, which leads to LVH. The loss of compliance also increases the speed of the arterial pulse wave, which travels from the aortic root to the periphery and is reflected back to the heart during diastole. Normally, the reflected pulse wave increases pressure in the coronary sinus and thereby improves myocardial perfusion. However, when the pulse wave velocity (PWV) is increased, the reflected wave arrives at the heart before the end of systole, which increases the systolic blood pressure and decreases the diastolic blood pressure. This increase in pulse pressure further contributes to cardiac work load. Perfusion of the myocardium during diastole is decreased, which contributes to myocardial ischemia. Increased PWV is an independent predictor of cardiovascular mortality in adults with ESRD.25
Intimal calcification is associated with classic risk factors for atherosclerosis, such as age, diabetes mellitus, smoking, high LDL cholesterol levels and inflammation, whereas medial calcification is strongly associated with ESRD-specific factors, such as hypertension, long-term dialysis and use of calcium-containing phosphate binders.24 High serum phosphate levels can induce both intimal and medial calcification.24 A high serum phosphate level has been associated with increased risks of mortality and cardiac disease in adults, even in the absence of renal disease.26 High extracellular phosphate levels induce vascular smooth muscle cells to differentiate into an osteoblastic phenotype that shows increased expression of osteopontin and alkaline phosphatase.27, 28 'Uremic' serum might also, however, induce arterial calcification and osteoblastic differentiation of vascular cells in the absence of high phosphate levels.29 The factors in uremic serum that could be responsible for medial calcification and osteoblastic differentiation of vascular cells are oxidative stress, parathyroid hormone (PTH) fragments and vitamin D.12 PTH itself prevents vascular calcification, but elevated levels of PTH fragments, many of which are competitive inhibitors of the PTH receptor, might increase vascular calcification.29
Intimal disease, in the form of irregularly increased carotid intima–media thickness (cIMT), can be detected at an early stage by ultrasonography. Medial disease is characterized by an increase in PWV, reduced carotid wall elasticity (as assessed by M-mode ultrasonography) and a less pronounced increase in cIMT with a more regular pattern of thickening than is seen in intimal disease. In 247 healthy adolescents aged 10–20 years, systolic blood pressure was the main predictor of vascular stiffening, whereas cIMT was only increased in individuals with high pulse pressure and high body weight, a well-established risk factor for atherosclerosis.30 Increased cIMT and carotid wall stiffening (as detected by ultrasonography), as well as increased carotid–femoral PWV (as detected by applanation tonometry), all strongly predict cardiovascular events in adults.31 Increased arterial wall stiffening, as assessed by either PWV or carotid ultrasonography, is highly prevalent in children on chronic dialysis.32, 33 Data on cIMT in children with ESRD are conflicting; most studies have reported significantly increased cIMT in children and young adults on RRT compared with that in healthy controls, but some studies found small or almost no increases in cIMT.33, 34, 35, 36
In one study of children who had CKD or were on dialysis, cIMT was higher than in children who had a renal transplant.35 cIMT was independently associated with factors related to calcium–phosphate metabolism.35 On follow-up of the same children, cIMT increased progressively, but decreased after transplantation.37 Shroff et al.33 found that cIMT and aortic PWV were normal in children on dialysis who had well controlled serum phosphate and intact PTH levels, but were increased in those with high serum phosphate and intact PTH levels. Although occlusive atherosclerotic lesions are rare in children, extensive coronary calcifications have been observed in adolescents and young adults with ESRD.7, 38 Civilibal et al.39 found coronary calcification by use of CT in 8 of 53 (15%) patients with ESRD aged 11–21 years, of whom 6 were on dialysis and 2 had a renal graft. All the previously mentioned risk factors—particularly high serum phosphate level and high calcium and calcitriol intake—were considerably more prevalent in patients with coronary calcification than in those without.
Cardiac changes
LVH is caused by several conditions commonly found in patients with CKD, including anemia, hypertension, hypercirculation due to arteriovenous fistulae and increased arterial stiffness. Moreover, the structure of the myocardium itself is altered in patients with ESRD, in a manner that is characterized by an increase in interstitial fibrosis and a reduction in myocardial capillary density.40, 41 These alterations influence the conductive properties of the myocardium and exacerbate the risk of life-threatening arrhythmias. LVH can be either concentric (as a result of pressure overload from increased systolic blood pressure and arterial wall stiffening) or eccentric (as a result of volume overload and anemia). Both volume and pressure overload induce myocardial remodeling by triggering events such as the release of angiotensin II and aldosterone, activation of the sympathetic nervous system and inflammation. This remodeling leads to predominantly diastolic dysfunction and ultimately to systolic dysfunction, cardiac failure and arrhythmia.
Circumstantial evidence of myocardial remodeling has been found in children with ESRD. Prolongation of the QT interval can occur during hemodialysis in children, even without demonstrable electrolyte disturbances.42 Children with CKD have notably longer corrected QT intervals (QTc) than healthy children and the duration of the QTc positively correlates with the duration of renal failure.43 However, QTc duration is not related to left ventricular mass index,43 which implies that alterations of the myocardial tissue might contribute to the increased risk of sudden cardiac death, independently of LVH.
LVH, as assessed by ultrasonography, is a major risk factor for acute cardiac death in adults with ESRD, and the most common cardiac abnormality in pediatric patients with ESRD.44, 45, 46 Concentric LVH can be found as early as stage 2–4 CKD in children.47 The incidence of LVH at the onset of RRT varies from 54% to 82% in children.45, 48, 49 LVH improves in most children after transplantation, provided that their blood pressure is well controlled.49, 50
As in studies of adults with renal failure, most studies in children with renal failure show strong associations between hypertension and LVH. Other studies of the relationship between blood pressure and LVH in children with conservatively treated renal failure report conflicting results. In the ESCAPE trial, no relationship between blood pressure and LVH was found,51 but longitudinal studies that involved ambulatory blood pressure measurement did show a relationship.48, 52 Therapeutic goals for blood pressure control, which should prevent LVH, are not always attained. Data from the Chronic Kidney Disease in Children study (CKiD), which enrolled patients with glomerular filtration rates of 30–90 ml/min/1.73 m2, indicated that the overall prevalence of hypertension was 54% and that 48% of patients who were being treated for hypertension did not have adequately controlled blood pressure.53
Therapeutic strategies
Angiotensin-converting-enzyme (ACE) inhibitors and angiotensin receptor blockers (ARBs) can reduce endothelial damage and are cardioprotective in adults with CKD. Data from the CKiD study suggest that use of ACE inhibitors and ARBs is associated with improved blood pressure control in children with CKD.51 Supplementation of L-arginine, the substrate for production of nitric oxide by eNOS, might improve endothelial function. However, Bennett-Richards et al.54 found that 4 weeks of oral L-arginine supplementation had no effect on endothelial function in children with CKD compared with placebo, despite eliciting a substantial rise in plasma L-arginine levels. Although the same investigators did demonstrate an improvement in flow-mediated dilatation and a decrease in plasma homocysteine levels after 8 weeks of oral folic acid supplementation in children with CKD, flow-mediated dilatation was not different between the actively treated group and the placebo group at the same timepoint.55
As in adults, high serum levels of phosphate and of calcium–phosphate product are thought to be among the most important risk factors for vascular and cardiac valve calcification in children with ESRD. This observation implies that vigorous treatment of hyperphosphatemia with (preferably calcium-free) phosphate binders is essential to prevent cardiovascular calcification. Although the evidence is less than convincing, some data suggest that effective treatment of hyperparathyroidism also prevents cardiovascular calcification.33
Transplantation is associated with reductions in arterial stiffening and LVH, and it sharply reduces the risk of cardiac death. Becker-Cohen et al.49 found that the prevalence of LVH decreased from 54% to 8% after transplantation in 13 children. Nevertheless, cardiovascular disease remains highly prevalent after transplantation in children with ESRD compared with in healthy children. In the LERIC study, 75% of all men with juvenile ESRD who were aged 20–40 years had functioning renal allografts, and nearly 50% had apparent hypertension-associated LVH.50 Briese et al.56 found increased vascular stiffness and signs of endothelial dysfunction in a study of 36 children with renal transplants. Bilginer et al.57 found that cIMT was increased in 24 renal transplant recipients aged 9–24 years compared with that in healthy controls; the extent of cIMT increase was associated with length of time on dialysis and the mean historical level of calcium–phosphate product.
Left ventricular mass index improves in children over a period of 2 years after commencing hemodialysis, in parallel with improvements in blood pressure and volume control.45 Intensified hemodialysis regimens, especially quotidian nocturnal hemodialysis, might eliminate the majority of risk factors for cardiovascular disease, by reducing serum ADMA, phosphate, and intact PTH levels, increasing HDL cholesterol levels and normalizing mean blood pressure. In adults, nocturnal hemodialysis is associated with disappearance of LVH, improvement of endothelial function, restoration of impaired peripheral vascular flow and cessation of vascular calcification.58 The very limited data on this modality in children suggest the same beneficial effects.59, 60 In our experience, nearly all children feel that the dramatic increase in well-being derived from daily dialysis, especially under a nocturnal regimen, by far compensates for the increased social burden. In our opinion, such intensified hemodialysis regimens are mandatory for children who cannot undergo transplantation immediately.
Conclusions
The risk of cardiovascular death is extremely high in children with renal disease, and sudden cardiac death is the main cause of cardiac death in these individuals. Classic risk factors for atherosclerosis are less prevalent in children with ESRD than in adults, which explains why medial calcification is more apparent than intimal disease in young patients. Several modifiable risk factors, including hyperphosphatemia, hyperparathyroidism, anemia and hypertension, independently predict the presence of cardiovascular abnormalities in this setting (Table 2). Increased awareness of the importance of controlling these nontraditional risk factors, among physicians who care for children with CKD or ESRD could improve the survival of these patients. New insights into the roles of inflammation and microvascular function in cardiovascular disease, the factors involved in the prevention of metastatic calcification, the activities of circulating EPCs, and novel treatment strategies for renal osteodystrophy might also contribute to improved survival. Large, prospective, international studies are warranted to address these issues, and to provide improved treatment options for this vulnerable population in the future.
Review criteria
Material for this article was retrieved by searching PubMed using the terms (alone or in various combinations) "kidney failure, chronic" OR "end-stage" OR "chronic" AND "kidney" OR "renal" AND "disease" OR "failure" AND "myocardium" OR "cardiomyopathy" OR "cardiac" AND "hypertrophy" OR "function" OR "dysfunction" OR "arteriosclerosis" OR "arterial" AND "compliance" OR "stiffness" OR "pulse wave velocity" OR "intima–media thickness" OR "endothelium" AND "function" OR "dysfunction" OR "repair". Results were limited to studies that focussed on individuals aged less than 18 years.
