 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
,
,#,#,¶¶,||||,**
* Department of Cardiology
Division of Nephrology, Department of Medicine
Department of Laboratory Medicine, Children’s Hospital, Boston, Massachusetts
¶ Division of Biostatistics, Dana Farber Cancer Institute, Boston, Massachusetts
|| Department of Biometry and Epidemiology, Medical University of South Carolina, Charleston, South Carolina
# Department of Pediatrics
** Department of Pathology, Harvard Medical School, Boston, Massachusetts
Division of Pediatric Cardiology, Golisano Children’s Hospital at Strong and University of Rochester Medical Center
Department of Pediatrics, University of Rochester School of Medicine and Dentistry, Rochester, New York
¶¶ Division of Nephrology, Vanderbilt Children’s Hospital
|||| Department of Pediatrics, Vanderbilt University School of Medicine, Nashville, Tennessee
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective. To determine the relationship between uremia and cTnT levels.
Design. Echocardiograms and blood chemistry results were obtained from 50 pediatric patients with chronic renal failure and without clinical heart disease. Levels of cTnT were tested for correlation with cardiac dysfunction. In multivariate analysis, biochemical aspects of renal disease and its treatment were tested for correlation with cardiac dysfunction.
Results. Forty-nine patients had cardiovascular abnormalities, including increased left ventricular function and mass, elevated heart rate and blood pressure, and reduced LV afterload. LV contractility was inversely correlated with cTnT level (r = –0.36). Higher cTnT also correlated with higher serum creatine kinase-MB mass, lower serum parathyroid hormone, higher blood urea nitrogen and bicarbonate levels, and the use of diuretics, but not with higher cardiac troponin I. Left ventricular contractility was inversely related to serum creatinine, phosphorus, and the use of ß-blockers.
Conclusions. Elevated cTnT levels are not artifactual, but are genuine indicators of cardiomyocyte damage. Cardiac damage, indicated by either elevated cTnT or low LV contractility, is related to uremia, deranged calcium and phosphorus metabolism, and bicarbonate levels. Serum cTnT and LV contractility identify subclinical cardiac damage that could be treated to hopefully reduce cardiovascular morbidity and mortality in this high-risk population.
Key Words: chronic renal failure • hemodialysis • pediatrics • troponin • cardiomyopathy • myocardial dysfunction • child • uremia • chronic renal insufficiency
Abbreviations: CRF, chronic renal failure • CK-MB, serum creatine kinase-MB mass • cTnT, serum cardiac troponin T • GFR, glomerular filtration rate • CRI, chronic renal insufficiency • LV, left ventricular • BUN, blood urea nitrogen • PTH, parathyroid hormone • cTnI, serum cardiac troponin I
Cardiovascular disease is the leading cause of death in chronic renal failure (CRF) patients on dialysis, as well as a leading cause of morbidity.1 Rates of cardiovascular disease are up to 20 times higher in adult CRF patients than in the general population.2 Clinical and subclinical myocardial ischemia are common among CRF patients, both before and during dialysis.3,4 Earlier detection of cardiovascular abnormalities in these patients might allow earlier interventions to reduce morbidity and mortality.
Studies of cardiac markers in this population have detected low-level elevations of serum creatine kinase, serum creatine kinase-MB mass ([CK-MB]; the cardiomyocyte-specific fraction), CK-MB isoforms, and serum cardiac troponin T (cTnT).5–9 However, either these findings have not been attributed to myocardial injury (because of factors such as lack of assay specificity), or they have been attributed to injury caused by risk factors other than uremia. These risk factors have included advanced age, atherosclerosis, diabetes, human immunodeficiency virus infection, traumatic injury, and the use of alcohol, cigarettes, or drugs. In addition, some of these previous studies used clinical indicators that were not sensitive enough to detect cardiomyocyte health. Thus, it was not possible to determine if serum marker elevations indicated subclinical myocardial injury related to CRF. A pediatric population with CRF provides an ideal group for the study of cardiac status in uremia, because these patients are less likely to have these potentially confounding risk factors.
In this prospective study, we studied pediatric patients with CRF but without overt heart disease or a history of ischemia or heart disease. We tested elevated serum cardiac markers for correlation with sensitive echocardiographic measures of cardiac structure and function. We also investigated whether any detected cardiac injury was related to either uremia or treatments for uremia.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Renal Therapy
All hemodialysis patients were treated with high-efficiency dialysis using polysulfone dialyzers, bicarbonate-based dialysate, and controlled ultrafiltration. Dialyzers were reprocessed with formalin a maximum of 10 times. Dialysis adequacy was assessed monthly with formal urea kinetic modeling and KT/V, a standard index of dialysis efficiency that is calculated by comparing the patient’s predicted and actual clearance of dialyzable solutes. All patients had KT/V 
1.4, which exceeds National Kidney Foundation-Dialysis Outcome Quality Initiative recommendations.10 The peritoneal dialysis patients were all receiving continuous cycling peritoneal dialysis. Concomitant medications for all dialysis patients included calcium carbonate, calcitriol, and water-soluble vitamins including folate.
Echocardiographic Evaluation
All patients underwent a single study echocardiogram. In the hemodialysis patients, echocardiography was performed immediately before hemodialysis. Echocardiograms were interpreted by echocardiographers masked to the patient’s clinical history. The echocardiographic study consisted of a complete 2-dimensional and Doppler evaluation. The combined M-mode echocardiogram, phonocardiogram, and pulse tracings were digitized and analyzed as previously described11 to derive values for systolic performance, preload, afterload, and contractility. Both conventional endocardial and newer midwall indices of LV performance were measured. In the setting of LV hypertrophy, endocardial indices are artifactually elevated relative to midwall indices, and thus may fail to detect clinically important LV dysfunction found by midwall measurements.12–14 However, midwall indices may be less sensitive because of inaccuracy in wall thickness measurement. Echocardiographic results were compared with data from 296 healthy subjects who were studied according to the same protocol.11
Assays
In hemodialysis patients, serum was taken both before and after hemodialysis. In patients on peritoneal dialysis and CRI patients, serum was obtained at any convenient time. No patient had cardiogenic shock or rhabdomyolysis at the time of sampling. All samples were stored at –70°C for <60 days. Cardiac troponin T was measured by enzyme immunoassay (Elecsys Troponin T STAT) on an Elecsys 1010 system (Roche Diagnostics Corp, Indianapolis, IN). A value of 0.01 ng/mL or higher was considered elevated on the basis of our analytical validity testing. Cardiac troponin I was measured by the Opus Plus analyzer (Behring Diagnostics, Westwood, MA) and in a similar serum cardiac troponin I (cTnI) assay by the Stratus II analyzer (Dade International, Miami, FL). A value of 0.05 ng/mL or greater was considered elevated. The chemical markers listed in Table 2 were measured by standard laboratory methods.
|
of .05. All echocardiographic variables were expressed as z scores (normal deviates) relative to the age- or body surface area-appropriate distribution in healthy children as previously described.11 We used a step-up regression procedure (keeping values with P < .05 at each step) to examine the relationships between patient characteristics, serum biochemistry, and cardiac function z scores, controlling for other variables such as hypertension and tachycardia. We examined all the exploratory variables listed in Tables 1 through 3 as predictors. For these step-up regression models, partial correlation coefficients were used to quantify the strength of the association between serum biochemistry and cardiac z scores, controlling for other variables in the model, including cardiac variables such as hypertension. These partial correlation coefficients are given in Tables 4 through 6. Because this was an exploratory analysis, we did not adjust for multiple testing. Therefore, although highlighted significant associations appear meaningful, the P values should be interpreted cautiously.
|
|
|
|
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Patients on hemodialysis were significantly older than those with CRI and were more likely to be receiving erythropoietin and ß-adrenergic antagonist therapy (Table 1). They also had higher serum levels of creatinine and cTnT and lower levels of CK and bicarbonate (Table 2).
Serum Chemistry
Serum cTnT was elevated in 21 patients. Three of 20 (15%) CRI patients had cTnT elevations and 18 of 27 (67%) hemodialysis patients had cTnT elevations. Hemodialysis patients were significantly more likely to have cTnT elevations (P 
.0001). The median cTnT level was 0.013 ng/mL (range: 0–0.239 ng/mL) in hemodialysis patients. All hemodialysis patients with elevated predialysis levels of cTnT also had elevated postdialysis levels. There were no isolated postdialysis cTnT elevations. Pre- and postdialysis cTnT levels were closely correlated (Pearson’s r = 0.98, P < .001), as were predialysis and postdialysis serum CK-MB mass values (r = 0.45, P = .019), suggesting that CK-MB and cTnT are poorly dialyzable under the stated conditions.
Serum cTnI measured by the Behring assay was elevated in 6 patients, 11.1% (3 of 27) of patients on hemodialysis and 15% (3 of 20) of patients with CRI. Two of 3 patients on hemodialysis with elevated cTnI also had elevated cTnT. No patient with CRI who had elevated cTnI had elevated cTnT. No patient had elevated cTnI according to the Dade assay.
CK-MB was elevated (>5 ng/mL) in 6 of the 49 patients (12.2%) who agreed to be followed. In patients on hemodialysis, CK-MB was elevated in 14.8% (4/27) before dialysis and 11.1% (3/27) after dialysis, with no clear pattern of increase or decrease between pre- and postdialysis levels. Two of the patients with CK-MB elevation had cTnT elevation as well. No patient with an elevated CK-MB had elevated cTnI.
Echocardiography (Table 3)
Only 1 of the 50 patients had a normal echocardiogram. The echocardiographic findings based on endocardial measurements included hyperdynamic LV systolic performance with increased LV fractional shortening, reduced LV afterload related to increased LV wall thickness, an increased thickness-to-dimension ratio, and increased mass. Heart rate and diastolic blood pressure were significantly elevated as well. Height and weight at the time of echocardiography were significantly reduced for age and sex. Other anatomic findings by echocardiography included pericardial effusion in 9/50 (18%) patients, aortic regurgitation in 7/50 (14%) patients, and a dilated aortic root in 4/50 (8%) patients. No patient had regional wall motion abnormalities. In the CRI group, the means of 4 aspects of LV function were abnormal (mass, end-diastolic dimension, end-diastolic posterior wall thickness, and stress-shortening index). In the hemodialysis group, 5 measures were abnormal (mean LV mass, fractional shortening, end-diastolic posterior wall thickness, thickness-to-dimension ratio, and afterload).
Multivariate Analysis (Tables 4 Through 6)
Lower LV load-independent contractility was significantly correlated with a lower GFR and not receiving dialysis (Table 4). Lower LV preload-dependent contractility was significantly correlated with higher serum cTnT (Table 5). Serum cTnT correlated directly with CK-MB (Table 5). Serum cTnI did not correlate with any other variable.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
We next examined the patients with either impaired LV contractility or elevated cTnT, because both could be considered to be suffering from myocellular injury. In these myocellular-injury patients, we used regression analysis to determine which aspects of uremia or its treatment contributed to the injury. Overall, worsening renal disease (indicated by increased creatinine and decreased GFR) correlated with LV dysfunction. Tables 4 and 5 show that myocardial impairment or injury correlated with 2 classes of risk factors: uremic substances or markers, and derangements in phosphorus and PTH homeostasis. It also correlated with bicarbonate levels. Also, cTnT abnormalities were more common in patients on dialysis than in patients with CRI.
The echocardiographic findings included increased LV mass with an elevated thickness-to-dimension ratio, reduced afterload (wall stress), elevated endocardial function indices with normal midwall indices, and normal preload. In contrast to adults with CRF,15 this pediatric population had wall stress results implying adequate compensatory hypertrophy.
The cardiac status of children and young adults with CRF is characterized by: hyperdynamic endocardial LV systolic performance with normal contractility; reduced LV afterload related to increased LV wall thickness, mass, and thickness-to-dimension ratio; and an augmented hemodynamic status with tachycardia and hypertension. Cardiac sympathetic overactivity is consistent with these findings,11,16,17 which may ultimately lead to depressed LV contractility, resulting in increased cardiac morbidity and mortality. This study demonstrated significantly lower LV contractility in children taking ß-blockers, reflecting the ability of ß-blockers to reduce adrenergic cardiovascular effects.18
Depressed contractility (defined as levels below the 95% confidence interval for healthy children) was not found in this study. However, as contractility was inversely associated with cTnT levels, it is possible that cardiomyocyte damage is a cause of impaired contractility, and that cumulative injury may result in clinically detectable effects over time.
Worsening renal disease (increased creatinine and decreased GFR) correlated with LV dysfunction. In adult patients with CRF, higher rates of cardiac abnormalities are associated with lower delivered dialysis doses,19 and LV function improves after dialysis in children with CRF,17,20 which is found in this study as well. These observations suggest that dialysis improves cardiac status at a clinical level,21 but do not address the question of whether dialysis prevents subclinical damage. We found that cTnT abnormalities were more common in patients on dialysis than in patients with CRI, which could suggest that dialysis itself is associated with subclinical damage, or that worsening renal function in these patients is associated with both hemodialysis and additional cardiac abnormalities. Other studies have supported dialysis-associated cardiomyocyte injury.9 A decrease in CK and cTnT and an improvement in cardiac status after successful renal transplantation22,23 or nephrectomy24 suggests that one or more factors associated with uremia are cardiotoxic.
In all reports, more patients with CRF have detectable serum cTnT than cTnI,25,26 perhaps because there is a lack of mass standardization of cTnI assays and there is heterogeneneity in the cross-reactivities of antibodies to various troponin I forms.27 Our results confirm that serum cTnI is less sensitive to cardiac injury and suggest that measuring low-level cTnT elevations may help predict or allow the early initiation of preventive strategies to hopefully prevent or reduce cardiovascular morbidity and mortality in this population. Before this study, it was unclear whether cTnT elevations represented subclinical cardiac injury, extracardiac expression of cTnT, or cross-reactions of antibodies associated with CRF or subclinical myocardiocyte damage from cardiac risk factors unrelated to CRF. Elevations of cTnT in CRF are not explained by expression of skeletal muscle troponin T, but are correlated with increased cardiac risk.28,29 Among adults with CRF, median levels of cTnT are 0.21 ng/mL in nondiabetics on hemodialysis, 0.39 ng/mL in diabetics on hemodialysis, and 0.06 ng/mL in diabetics with normal creatinine.30,31 In contrast, the median level among our younger patients was 0 ng/mL (maximum, 0.70 ng/mL). This lower value is to be expected, because young patients have fewer nonuremic cardiac risk factors.
Our results furthermore suggest that serum cTnT may be used to detect minor, clinically occult, myocardial injury in this population, allowing for preventive or corrective strategies. Such an approach would be worthwhile because cardiovascular complications remain the most common cause of death in pediatric patients with CRF. In fact, the percentage of deaths from cardiovascular sequelae are as common in children with CRF as adults.32,33 In addition, young adults with childhood-onset CRF have a high prevalence of persistent arteriopathy associated with indicators of microinflammation, hyperparathyroidism, calcium-phosphate overload, and hyperhomocysteinemia, but not traditional atherogenic risk factors.34
Among our patients, low-level elevations of serum cTnT were common, suggesting subclinical cardiomyocyte injury in many patients on dialysis or with CRI. Although dialysis can reduce injury caused by uremic toxins, it may also increase the cardiotoxicity of other factors, such as calcium, phosphorus, PTH, and bicarbonate. Previous studies of cTnT and of other markers have not examined correlations with LV contractility. The specificity of cTnT elevations for detection of active myocardial damage is supported in this study by significant correlations with both CK-MB and LV contractility. In addition to the significant correlation between preload-dependent contractility and serum cTnT, our data also suggested that there may be a weak correlation between load-independent LV contractility and elevated serum cTnT (r = –.29, P = .099; data not shown). In adults with CRF, low-level elevations of cTnT appear to be predictive of subsequent clinically significant cardiovascular morbidity or mortality.29,35
Our data suggest that uremia contributes to heart disease, because patients with CRF who had lower GFR before beginning dialysis had lower contractility, and because higher concentrations of uremic substances, such as serum creatinine and BUN, correlated with lower contractility and cTnT elevations. Other researchers have found a relationship between cTnT and BUN or creatinine. The concentration-dependent cardiodepressant effects of uremic markers have been described in numerous experimental studies.36–38
As our data imply, abnormalities of calcium and phosphorus homeostasis may be risk factors for the active myocardial injury demonstrated by elevations of cTnT, and could affect the morbidity of patients with CRF. Although we have not explored the exact cause of myocardial injury, ischemia can cause myocardial injury and ischemia is a problem in adults with CRF. PTH is an independent factor influencing cardiac function in pediatric CRF.39 Low PTH levels are correlated with LV dilation in adults with CRF,3 just as lower PTH was correlated with elevated cTnT in our study. The experimental induction of acute hypercalcemia in patients with CRF increases serum-ionized calcium, decreases plasma intact PTH, and significantly impairs LV diastolic function compared with healthy controls.40 Lower PTH may be related to lower intracellular calcium concentration, which can occur during dialysis and is a strong correlate of ischemia,41 depressed contractility,42 and adverse outcomes.43 Impaired cardiac function in experimental renal failure is associated with abnormal cardiac energetics and increased susceptibility to ischemic damage.44 Disordered mitochondrial calcium usage may contribute to these derangements, resulting in impaired cytosolic calcium control, disturbed contractile function, and reduced energy reserves. The use of calcium channel antagonists in this study had a significant effect on LV fiber stress, a result that further demonstrates the effect of calcium on cardiac function.45 In this study, we also found an inverse correlation between phosphorus and both LV contractility and blood pressure. Increased phosphorus has been associated with accelerated formation of calcific valves and advanced coronary plaques, LV dilation, hypotension, bradycardia, and myocardial ultrastructural changes.3,18
Our study suggests a direct relationship between serum bicarbonate and cTnT. The value of bicarbonate use in patients with metabolic acidosis has been questioned. Bicarbonate has disadvantages with respect to hemodynamics, LV function, blood lactate, intracellular pH, oxygen delivery and consumption, and clinical outcome.46–48 Bicarbonate may worsen the local injury in patients with active cardiomyocyte injury.
Multiple cardiac risk factors exist in patients with CRF and may act in concert to produce a high incidence of cardiovascular disease.49 Identifying modifiable independent risk factors for subclinical cardiac disease may lead to clinical strategies to reduce heart disease and mortality. Our study suggests that both uremia and its therapy have cardiovascular risks. Even in the absence of diabetes or atherosclerotic disease, patients with CRF had cardiac injury that was related to factors that are potentially modifiable. Measuring cTnT and LV contractility could provide objective standards of cardiovascular status in patients with CRF and of their response to clinical interventions.50, 51
Although other studies of adults have established a relationship between inadequate dialysis (low KT/V) and cardiovascular morbidity and mortality,52 in this study all patients had adequate dialysis as assessed by KT/V and did not have clinically apparent myocardial disease. This suggests that a KT/V of 1.4 may not be sufficient to avoid myocardial injury or to serve as a predictor of cardiovascular morbidity and mortality for younger patients. Children have longer projected lifespans than older hemodialysis patients and are more likely to have time for the subtle abnormalities detected in this study to evolve into clinically apparent disease. Although a KT/V of 1.4 reflects adequate dialysis, it may not be optimal. The findings of this study suggest further investigation of the appropriate outcome measures in assessing dialysis adequacy in children. Adequate control of myocardial injury may become a part of the definition of adequacy of dialysis, along with an adequate dose of dialysis and adequate protein intake.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Address correspondence to Steven E. Lipshultz, MD, Division of Pediatric Cardiology, University of Rochester Medical Center, 601 Elmwood Ave, Box 631, Rochester, NY 14642. E-mail: steve_lipshultz{at}urmc.rochester.edu
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. L. Aschner and R. L. Poland Sodium Bicarbonate: Basically Useless Therapy Pediatrics, October 1, 2008; 122(4): 831 - 835. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Becker-Cohen, A. Nir, C. Rinat, S. Feinstein, N. Algur, B. Farber, and Y. Frishberg Risk Factors for Cardiovascular Disease in Children and Young Adults after Renal Transplantation Clin. J. Am. Soc. Nephrol., November 1, 2006; 1(6): 1284 - 1292. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Atar, M. E. Korkmaz, I. A. Atar, O. Gulmez, B. Ozin, H. Bozbas, T. Erol, A. Aydinalp, A. Yildirir, M. Yucel, et al. Effects of metoprolol therapy on cardiac troponin-I levels after elective percutaneous coronary interventions Eur. Heart J., March 1, 2006; 27(5): 547 - 552. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. E. Lipshultz, N. Rifai, V. M. Dalton, D. E. Levy, L. B. Silverman, S. R. Lipsitz, S. D. Colan, B. L. Asselin, R. D. Barr, L. A. Clavell, et al. The Effect of Dexrazoxane on Myocardial Injury in Doxorubicin-Treated Children with Acute Lymphoblastic Leukemia N. Engl. J. Med., July 8, 2004; 351(2): 145 - 153. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2003 American Academy of Pediatrics. All rights reserved. |
||
|
||
