Abstract
Cardiovascular disease is increasingly recognized as a life-limiting problem in young patients with chronic kidney disease, but there are few studies in children that describe its determinants. We studied the association of intact parathyroid hormone (iPTH) levels and their management on vascular structure and function in 85 children, ages 5–18 years, who had received dialysis for ≥6 months. Compared to controls, dialysis patients had increased carotid intima-media thickness and pulse-wave velocity. All vascular measures positively correlated with serum phosphorus levels, while carotid intima-media thickness and cardiac calcification score also correlated with iPTH levels. Patients with mean time-integrated iPTH levels less than twice the upper limit of normal (n = 41) had vascular measures that were comparable to age-matched controls, but those with iPTH levels greater than twice the upper limit of normal (n = 44) had greater carotid intima-media thickness, stiffer vessels, and increased cardiac calcification than controls. Patients with increased carotid intima-media thickness had stiffer vessels and a greater prevalence of cardiac calcification. There was a strong dose-dependent correlation between vitamin D and all vascular measures, and calcium intake from phosphate binders weakly correlated with carotid intima-media thickness. In conclusion, both iPTH level and dosage of vitamin D are associated with vascular damage and calcification in children on dialysis.
Cardiovascular disease is the most common cause of death in young adults with chronic kidney disease (CKD). 1 Altered morphologic properties of the large arteries can be seen from the second decade of life, 2–6 imparting a 700-fold increased risk for death over the age-adjusted normal population. 7 In recent years, attention has focused on the role of calcium (Ca), phosphate (PO4), and parathyroid hormone (PTH) as potentially modifiable risk factors in the development and progression of vascular disease. 8–10
Intact PTH (iPTH) levels of two to four times the upper limit of normal (ULN) are recommended for patients who are on dialysis 11,12 on the basis of bone histomorphometry studies in children 13 and adults. This is because as CKD progresses, continued stim-ulation of the parathyroid glands by high plasma PO4 and low Ca levels leads to parathyroid gland hypertrophy and re-setting of the calcium-sensing receptor so that higher than normal levels of ionized Ca are required to alter PTH secretion. This “skeletal resistance” to PTH results in low-turnover (adynamic) bone disease and an inability of bone to buffer fluxes in serum calcium, resulting in ectopic soft tissue calcification. However, it has to be remembered that high iPTH levels are per se a risk factor for vascular disease and soft tissue calcification. 10,14,15
At Great Ormond Street Hospital, we aim to prevent the escape of the parathyroid glands from normal control mechanisms by dietary and therapeutic intervention. We have shown that keeping iPTH levels in the normal range in stages 1 through 4 CKD and <2 ULN in children who are on dialysis maintains normal growth velocity 16 and bone mineral density, 17 but the presence of renal osteodystrophy was ubiquitous in children with CKD irrespective of iPTH levels. 18
These long-standing differences in management of secondary hyperparathyroidism among pediatric renal units in the United Kingdom have given us the opportunity to study the impact of a wide range of iPTH levels and their management on vascular structure and function. We hypothesized that maintaining iPTH levels <2 ULN throughout the course of CKD will prevent vascular damage and calcification in children who are on dialysis.
Results
Dialysis and Control Groups
Diagnoses of the 85 study children (45 boys) were dysplasia (n = 50), inherited nephropathies (n = 13), cystic kidney disease (n = 6), primary tubular disorders (n = 6), renovascular disorders (n = 4), malignancies (n = 3), and metabolic disorders (n = 3). Comparisons between control subjects and groups I and II are shown in Table 1.
Demographic, clinical, anthropometric, and biochemical characteristics of patients and control subjects a
Carotid intima-media thickness (cIMT) and brachioradial pulse-wave velocity (PWV) were significantly greater in the dialysis population than in the control group (Table 2). The duration of dialysis was associated with an increasing cIMT (r = 0.31, P = 0.04) but not with PWV or the presence of cardiac calcification. The time spent in stage 4 CKD, age at initiation of dialysis, dialysis modality, and preservation of residual renal function did not correlate with any vascular measures.
Comparison of carotid artery structure, vascular stiffness, and calcification scores between dialysis and control patients and between groups I and II
Vascular Measures and Calcification Score in Groups I and II
cIMT.
cIMT showed a strong linear correlation with iPTH (r = 0.71, P = 0.0001), PO4 (r = 0.51, P < 0.0001; Figure 1) and Ca × PO4 (r = 0.65, P < 0.0001). The cIMT in group I was comparable to the control group (0.39 ± 0.01 versus 0.38 ± 0.01 mm; P = 0.44) and significantly lower than in group II (0.58 ± 0.02; relative risk 3.7; P < 0.0001; Figure 2).
Correlation of cIMT with mean time-integrated serum PO4 level.
Comparison of cIMT levels among control subjects, group I, and group II. Group I, mean time-integrated serum iPTH ≤2 ULN; group II, mean time-integrated serum iPTH >2 ULN.
PWV.
Aortic PWV also showed a positive correlation with PO4 levels (r = 0.39, P = 0.03) and Ca × PO4 (r = 0.37, P = 0.018). Aortic PWV was greater in group II than in group I (8.63 ± 2.3 versus 5.81 ± 1.2 m/s; P = 0.03; Table 2). However, the brachioradial PWV did not correlate with any demographic or biochemical parameters.
Cardiac Calcification Scores.
The calcification score was associated with iPTH levels (r = 0.39, P = 0.03; Figure 3) and serum PO4 (r = 0.34, P = 0.03) but did not correlate with age, duration of CKD, or time on dialysis. Five (12%) patients in group I and 12 (27%) in group II (P = 0.004) had calcification (relative risk 2.3; Table 2). Calcification was graded 19 as minimal (Agatston score <10) in four, mild (score 11 to 100) in six, moderate (score 101 to 400) in five, and severe (score >400) in two. Moderate and severe grades of calcification were seen only in group II. There was no definite anatomic pattern of calcium deposition in the vessels or valves. Three (17%) of 17 patients who were younger than 10 yr had calcification as compared with 14 of 68 (20%) who were older than 10 yr.
Correlation of Agatston score for cardiac calcification with mean time-integrated iPTH levels.
Effect of PO4 Binders and Vitamin D Therapy on Vascular Measures
Eighteen (41%) patients in group II received sevelamer with or without calcium-based PO4 binders compared with only five (12%) in group I (Table 1). The group of patients on sevelamer were older (15.1 ± 4.1 yr), had a longer dialysis vintage (3.0 ± 1.9 yr) and were predominantly on hemodialysis (73%) as compared with the overall cohort. The mean elemental Ca intake from the prescribed dosage of PO4 binders was marginally greater in group II than in group I (Table 1). Although the elemental Ca intake from PO4 binders did not correlate with any of the vascular measures in the overall cohort, after exclusion of patients who received sevelamer, the IMT did show a weak correlation with calcium-containing PO4 binder dosage (r = 0.19; P = 0.054).
The vitamin D dosage showed a strong dosage-dependent correlation with cIMT (r = 0.65, P < 0.001), aortic PWV (r = 0.17, P = 0.03), and calcification score (r = 0.28, P = 0.02; Table 3, Figure 4). Patients with calcification (n = 17) received a 2.8-fold higher vitamin D dosage than those without calcification (21.9 ± 8.9 versus 53.7 ± 13.4 μg/kg, P = 0.0001).
Correlation of cIMT with the mean calcitriol dosage.
Multiple regression analysis for independent predictors of cIMT, aortic PWV, and calcification score
Correlations between Vascular Measures
cIMT was associated with aortic PWV (r = 0.41, P < 0.0001) and the calcification score (r = 0.32, P = 0.002): Patients with calcification (n = 17) had significantly higher cIMT than those without calcification (0.55 ± 0.11 versus 0.45 ± 0.12 mm; P = 0.004; Figure 5).
Correlation of Agatston score for cardiac calcification with cIMT.
Predictors of cIMT, Aortic PWV, and Calcification Score
On multiple regression analysis, the vitamin D dosage was the strongest predictor of cIMT, aortic PWV, and calcification, whereas iPTH levels were an independent predictor of cIMT and calcification (Table 3).
Discussion
This is the largest pediatric study in the youngest cohort of CKD patients on dialysis that describes vascular changes and the impact of iPTH control and vitamin D treatment on these. We show that both hyperparathyroidism and its management with vitamin D have an impact on structural and functional vascular changes that begin as early as the first decade of life.
Uremia is a vasculopathic process. Children provide an ideal opportunity to study the uremic influences on the arterial wall because they rarely have risk factors such as diabetes, dyslipidemia, and hypertension that are prevalent in adults. 14,20 The patients in our study were carefully selected so that they were free of such confounders, and capitalizing on detailed serial biochemistry over ≥3.5 yr, we were able to demonstrate the impact of iPTH and its management on the vascular phenotype. Whereas PO4 has been shown to be an independent risk factor for cardiovascular disease, 8 including increased IMT, 4–6 vessel stiffness, 3,21 and left ventricular hypertrophy, 5 PTH per se may contribute to vascular injury via mechanisms other than its effect on Ca-PO4 homeostasis. It would be impossible to extricate the individual effects of PO4 and PTH and the medications used in their regulation. PTH may mediate vascular damage by playing a permissive role in arteriolar wall thickening and myocardial interstitial fibrosis, 10 increasing triglycerides and LDL cholesterol and contributing to chronic hypertension; progression of these vascular changes is reduced after parathyroidectomy. 22 Thus, although it is now widely accepted practice to aim for PO4 levels within the normal range, the optimal level for iPTH is as yet unclear. Results of this hypothesis-generating study will allow for a prospective randomized trial to evaluate the cardiovascular benefits of maintaining iPTH levels <2 ULN using the lowest possible dosage of vitamin D.
The long-term consequences of vascular damage are particularly important in children, who have a lifetime of renal replacement therapy ahead of them. Studies in adults with CKD have shown that approximately 65% have coronary calcification at the start of dialysis, 23 suggesting that prevention of secondary hyperparathyroidism is in fact key to the prevention of vascular damage and calcification. We and others have shown that endothelial dysfunction 24 and vascular damage 2–6 begin early in the course of CKD. The vascular damage is only partially reversible after transplantation 25 and use of lipid-lowering agents, 26 folate, 27 or arginine supplementation 28 has little effect. Goodman et al. 29 and others 23 have shown that once a nidus of calcification forms in the soft tissues, “calcium begets calcium” such that patients with preexisting calcification are at greatest risk for accelerated calcification. Thus, the prevention of secondary hyperparathyroidism from the earliest stages of CKD is key to preventing the development and progression of vascular calcification.
In young adults with childhood-onset CKD, Groothoff et al. 30 reported a significant increase in arterial stiffness but normal IMT, whereas a similar study by Oh et al. 31 showed increased IMT and calcification in 92% of his cohort. However, these studies in young adults can support only speculations of the potential changes in children who are on dialysis because it is possible that uremia multiplies the natural age-related vascular damage. Evidence of vascular changes in children who are on dialysis has come from observational studies that have shown increased IMT, 4–6 stiffer vessels, 3 and calcification, 2,29,32 and linked these with PO4 levels and Ca × PO4; however, patients in these studies were older than in our cohort and often had comorbidity, and the small patient numbers and widely variable duration of CKD and time on dialysis may have resulted in confounders in their analyses.
The management of secondary hyperphosphatemia with use of calcium-based PO4 binders has been under considerable debate with concerns that the Ca intake from PO4 binders results in calcium overload and ectopic soft tissue calcification. 4,11,12,14,29 This study was not designed to determine the vascular effects of different PO4 binders; sevelamer was used only as a second-line agent in our patients with persistently high iPTH levels and hypercalcemia. Goodman et al., 29 Litwin et al., 4 and Briese et al. 33 showed a positive correlation between the cumulative PO4 binder dosage and coronary calcification or cIMT, and in our study, IMT showed a weak correlation with PO4 binder dosage approaching statistical significance. However, adolescents and young adults are notoriously noncompliant with PO4 binder medication, making it difficult to assume that the prescribed dosage of PO4 binder is indeed what the patient consistently receives.
In our study, the vitamin D dosage was the most important predictor of increased arterial thickness, stiffness, and calcification. Although the prescribed dosage of vitamin D is also a surrogate marker of the severity of hyperparathyroidism, the vitamin D dosage predicted vascular damage independent of iPTH levels. The role of vitamin D in the pathogenesis of vascular calcification has been shown in other observational studies 4,5 as well as in ex vivo 34 and in vitro models. 35 Although the primary role of vitamin D is to increase the gastrointestinal absorption of calcium, it also significantly increases PO4 absorption. Moreover, vitamin D acts on the vascular smooth muscle cells via the vitamin D receptor and can induce proliferation and osteoblastic differentiation of these cells. 35
We found a significant positive correlation among cIMT, aortic stiffness, and presence of cardiac calcification, suggesting that the vascular damage is widespread, involving both large muscular arteries such as the carotids and elastic vessels such as the aorta. Unlike aortic PWV, the brachioradial PWV, although increased, did not correlate with any demographic or biochemical parameter or with other vascular measures. It is known that vascular damage in the aorta begins earlier than in the brachioradial vasculature, 21 suggesting that the lack of correlations with brachioradial changes may in fact be the result of dissociation in time rather than a difference in underlying pathology, as discussed in one other study. 36 A positive correlation between cIMT and calcification may suggest that deposition of Ca-PO4 crystals in the arterial media may be at least partly responsible for the increased cIMT. Carotid artery ultrasound, a cheap, easily available, highly reproducible, and noninvasive test to measure cIMT, may reliably substitute other methods for detection, monitoring, and prognostication of vascular damage in dialysis patients.
On multiple regression analysis, the age at study or time spent in CKD or on dialysis did not show any significant correlation with the vascular measures. Although increased vascular damage with age and dialysis vintage has been reported, 4,29–31 this may in fact be the result of prolonged hyperparathyroidism and its consequences on the vasculature. Goodman et al. 29 showed calcification only in patients who were older than 20 yr, but the group with calcification had significantly higher PO4 and Ca × PO4 than those without calcification. Subsequent studies by Eifinger et al. 32 and Civilibal et al. 2 documented coronary calcification in pediatric dialysis patients. In our cohort, age did not correlate with any vascular measure: The youngest patient with calcification was 5.8 yr.
Although we were unable to perform a randomized study, the groups were well matched for cardiovascular risk factors. The iPTH assay varied among participating centers, and the interassay variability of various PTH assays must be recognized. 37 Because of small patient numbers, our study lacked the power to show any potential differences in vascular measures on the basis of dialysis modality. Pediatric dialysis populations are currently limited because of a high transplantation rate; in the largest comparable study, 4 only 37 children were on dialysis.
In conclusion, we have shown that both the iPTH level and the vitamin D dosage are significant and independent predictors of vascular damage and calcification in children who are on dialysis.
Concise Methods
Study Design
Eighty-five children were recruited from four children’s renal units in the United Kingdom. Children who were aged 5 to 18 yr, had received dialysis for at least the preceding 6 mo, and had been in stage 4 CKD (GFR ≤30 ml/min per 1.73 m2) for ≥3 yr were included. Patients with underlying vasculitis, diabetes, uncontrolled hyperlipidemia (serum cholesterol ≥5 mMol/L despite statin therapy), uncontrolled hypertension (systolic BP ≥95th centile for age despite therapy), and smokers were excluded. Because vascular measures are age dependent and normative data are available only for children who were older than 10 yr, 38 we compared our patient group with 40 age- and gender-matched healthy control subjects who were studied contemporaneously.
A total of 103 eligible patients were identified, and 97 agreed to participate. Seven were excluded because of criteria identified at the first visit or unwillingness to undergo scans. Vascular measures were performed on 90 dialysis patients and 40 control subjects (five patients’ scans were of poor quality and excluded from analysis). Informed written consent was obtained from all parents or caregivers and from children when appropriate. The study was approved by a multicenter research ethics committee.
CKD-specific cardiovascular risk factors were measured from the start of stage 4 CKD, giving a minimum of 3.5 yr of cumulative data for all children. Serum Ca, PO4, and iPTH levels and the dosages of elemental calcium intake from phosphate binders and vitamin D were recorded at monthly intervals. Because the iPTH assay varied among centers, iPTH was expressed as multiples of the ULN. In control subjects, biochemical values from a single blood test at the time of the scans were used. Patients were divided into to groups on the basis of mean time-integrated serum iPTH levels: Group I with serum iPTH <2 ULN and group II with iPTH >2 ULN. Patients in group II included those for whom the unit policy was to aim for iPTH levels >2 ULN and also those with noncompliance to treatment despite an intention to keep iPTH levels <2 ULN. In all centers, calcium-based phosphate binders were used for management of hyperphosphatemia, with sevelamer introduced only for those with hypercalcemia and persistently high iPTH levels. Alphacalcidol (1-α hydroxy-vitamin D3) was used in all patients, titrating the dosage against iPTH levels, with temporary discontinuation when hypercalcemia resulted. All children received regular dietary advice.
All patients underwent a carotid artery ultrasound scan to measure cIMT, applanation tonometry for aortic and brachioradial PWV, and a multislice computed tomography (CT) scan to look for coronary artery and valvular calcification. In the control group, aortic PWV was not measured because permission was not granted for femoral artery measures, and CT was not undertaken because of radiation concerns.
All vascular scans were performed at Great Ormond Street Hospital after a midweek session of hemodialysis or overnight cycling peritoneal dialysis. Scans were performed by two trained operators who were blinded to the patients’ clinical and biochemical details. Intraobserver and interobserver reproducibility of measurements (determined by measurement of cIMT and PWV twice, 2 wk apart in 10 randomly selected children) was 1.9 and 2.2 μm, respectively, for cIMT and 0.11 and 0.23 m/s, respectively, for PWV.
Carotid Ultrasonography for cIMT
B-mode ultrasound of both common carotid arteries was performed (Vivid 7; GE Medical, Horton, Norway). On a longitudinal image of the vessel, 1 to 2 cm proximal to the carotid bulb, the optimal image was acquired on the R wave of the electrocardiogram (ECG) and frozen in diastole, and the distance between the leading edge of the lumen-intima interface and the media-adventitia interface on the far wall of the artery was measured using electronic calipers.
Applanation Tonometry for PWV
Applanation tonometry was performed with a micromanometer (SPC-301; Millar Instruments, Houston, TX) in the carotid, radial, and femoral arteries. An ECG gated signal and standard anthropometric measurements were used. PWV was computed from contour analysis of 10 consecutive pressure waveforms recorded consecutively in the carotid-radial and carotid-femoral arteries to give the brachioradial and aortic PWV, respectively.
Multislice-Spiral CT for Cardiac Imaging
A 16-slice spiral CT scan was performed (Somatom Sensation 16; Siemens Medical Solutions, Erlangen, Germany) using the standard calcium-scoring protocol (12-slice acquisition, 120 kV, 30 mAs, nominal slice width 3.0 mm, slice collimation 1.5 mm, gantry rotation time 0.42 s, and table feed 5.6 mm/rotation) and prospective ECG triggering. The Agatston score was determined for coronary arteries, cardiac valves, and aorta using standard methods. 39
Statistical Analyses
Results are presented as means ± SD. All data were analyzed in a linear manner and then between the two groups of dialysis patients. The t test, Mann-Whitney U test, or Fisher exact t test was used as appropriate. From univariate analysis, variables that were associated with vascular measures with P < 0.15 were entered into a stepwise multiple regression analysis. P ≤ 0.05 was considered statistically significant. Statistical analyses were performed using SPSS 12.0.1 (SPSS, Chicago, IL).
Disclosures
None.
Acknowledgments
This study was sponsored by the Kidney Research Aid Fund, and funding for CT scans was obtained from the Eyck and Strutt scholarship from the British Medical Association.
This study was presented at the European Society of Pediatric Nephrology, October 2006 (Palermo, Sicily), and won the award for the best oral communication.
We gratefully acknowledge Dr. Atul Singhal, Childhood Nutrition Research Centre, for contributing data on the control subjects; Dr. Mike Rubens, Department of Medical Imaging, Royal Brompton & Harefield NHS Trust, for advice on interpreting cardiac CT scans; and Cathy Mosby and Julia Holding (Clinical Nurse Specialists, St. James’s Hospital) and Donna Hilton and David Cooper (Nottingham City Hospital) for help with recruitment.
Footnotes
Published online ahead of print. Publication date available at www.jasn.org.
- © 2007 American Society of Nephrology
References
In this issue
Jump to section
More in this TOC Section
Cited By...
- Effects of Hemodiafiltration versus Conventional Hemodialysis in Children with ESKD: The HDF, Heart and Height Study
- Risk factors and determinants of carotid intima-media thickness in children: protocol for a systematic review and meta-analysis
- Decreased Arterial Elasticity in Children With Nondialysis Chronic Kidney Disease Is Related to Blood Pressure and Not to Glomerular Filtration Rate
- HDL in Children with CKD Promotes Endothelial Dysfunction and an Abnormal Vascular Phenotype
- Mechanistic Insights into Vascular Calcification in CKD
- Cardiovascular Disease in Children with Chronic Kidney Disease
- Ergocalciferol Supplementation in Children with CKD Delays the Onset of Secondary Hyperparathyroidism: A Randomized Trial
- Klotho Deficiency Causes Vascular Calcification in Chronic Kidney Disease
- The Cardiovascular Comorbidity in Children with Chronic Kidney Disease (4C) Study: Objectives, Design, and Methodology
- Vascular calcification in chronic kidney disease
- Chronic Mineral Dysregulation Promotes Vascular Smooth Muscle Cell Adaptation and Extracellular Matrix Calcification
- Pediatric Myocardial Stunning Underscores the Cardiac Toxicity of Conventional Hemodialysis Treatments
- MONITORING CARDIOVASCULAR RISK FACTORS IN CHILDREN ON DIALYSIS
- Dialysis Accelerates Medial Vascular Calcification in Part by Triggering Smooth Muscle Cell Apoptosis
- A Bimodal Association of Vitamin D Levels and Vascular Disease in Children on Dialysis