|Home | About | Journals | Submit | Contact Us | Français|
Children with advanced chronic kidney disease (CKD) frequently develop left ventricular (LV) hypertrophy. The extent of hypertrophy that results in cardiac dysfunction is unknown. Systolic function, routinely determined by ejection fraction (EF), is usually preserved in these patients. However, a decrease in EF represents an advanced cardiac dysfunction. We used cardiac magnetic resonance (CMR) and phosphorus-31 MR spectroscopy (31P MRS) to assess markers of cardiac dysfunction in young CKD patients.
Ten dialysis and ten post-transplant patients completed the study. The outcomes were peak LV myocardial circumferential strain (Ecc); myocardial T2 relaxation time and full width at half maximum (FWHM) of T2 distribution; and phosphocreatinine/adenosine triphosphate (PCr/ATP) to measure muscle energy metabolism. Healthy controls were used for comparison.
All patients had normal EF; nine (45%) had low Ecc. Ecc was lower in dialysis versus transplant (p<0.0001) patients and inversely correlated with LV mass index, r=−0.47, p=0.04. Patients had higher T2 (p=0.056) and FWHM (p=0.01) than controls. T2 levels were positively correlated with LVM index (r=0.46, p=0.04). PCr/ATP was lower in patients than in controls (p=0.02).
Young patients with advanced CKD and normal EF have early cardiac changes. Association of these abnormalities with increased left ventricular mass (LVM) index suggests development of maladaptive hypertrophy.
In adults, a strong association between chronic kidney disease (CKD) and cardiovascular disease (CVD), termed cardiorenal syndrome, is well known . The cardiorenal syndrome frequently manifests as left ventricular hypertrophy (LVH) coupled with systolic dysfunction. These abnormalities have been recognized as strong predictors for future CVD mortality in adult CKD patients [2–7]. Children and young adults with advanced CKD also develop LVH, which is frequently severe [8, 9]. Using echocardiography, we previously showed that LVH is associated with decreased diastolic function in children on maintenance dialysis and after kidney transplantation [10, 11]. In contrast, systolic function as measured by ejection fraction (EF) is usually preserved in these young patients [12–14]. When there is a decrease in EF, it represents advanced cardiac dysfunction. It is not known whether LVH in these patients produces functional abnormalities not detected by routine echocardiographic evaluation.
The aim of this study was to evaluate sensitive markers of cardiac structure, function, and metabolism and examine their association with LVH in children and young adults on maintenance dialysis and after kidney transplantation. We used cardiac magnetic resonance (CMR) and MR spectroscopy (MRS) for the study. The biomarkers included peak left ventricular (LV) myocardial circumferential strain (Ecc) to assess regional LV function, T2 relaxation time to quantify myocardial structural composition, and phosphocreatinine/adenosine triphosphate (PCr/ATP) ratio from phosphorus-31 MRS (31P MRS) to assess muscle energy metabolism. We hypothesized that early abnormalities of cardiac function were related to the severity of cardiac hypertrophy.
Twenty patients (ten on maintenance dialysis: six hemodialysis and four peritoneal dialysis; ten with renal transplants) were included and studied cross-sectionally. Twenty-four age-matched controls were used to compare LV structure, EF, and strain: 17 to compare T2 relaxation time and eight to compare energy metabolism. Inclusion criteria were:
Exclusion criteria were:
Healthy children and young adults were recruited from the personnel or the families of personnel at Cincinnati Children's Hospital. The institutional review board of Cincinnati Children's Hospital Medical Center approved the study, and informed consent was obtained for each study patient. Medical records were reviewed for age, sex, race, CKD etiology, and duration of renal failure or dialysis. Length of time following kidney transplantation was also recorded. All patients had a history and physical examination performed. Clinical and laboratory data were collected on the day of the CMR evaluation, including weight, height, systolic (SBP) and diastolic (DBP) blood pressure, serum creatinine, calcium, phosphorus, hemoglobin, and fasting serum lipids. Uncontrolled hypertension was defined as SBP or DBP above the age-, sex-, and height-specific 95th percentiles . The transplant allograft function was estimated by the Schwartz formula  in pediatric patients and by Modification of Diet in Renal Disease (MDRD) formula in young adults . Hemodialysis patients received dialysis treatment three times per week for 3.5–4.0 h in each session. Average ultrafiltration during dialysis, weight gain between dialysis treatments, and residual renal function were recorded. Dry weight was defined as body weight below which hypotension or muscle cramps occur. Children on peritoneal dialysis had daily treatment using continuous cycling peritoneal dialysis (CCPD). All medications (dose/day), including antihypertensives, lipid-lowering medications, and others, were recorded.
All study participants underwent CMR imaging and 31P MRS on a Philips 3.0 Tesla Achieva X-Series system (Philips Medical Systems, Best, The Netherlands) with maximum gradient strength of 40 mT/m at a 200-mT/m/ms slew rate. For cardiac imaging, the 32-channel phased-array coil was used. For 31P MRS, a 14-cm linearly polarized transmit and receive 31P coil was used. No intravenously administered contrast or sedation was used. Because energy metabolism can be affected by exercise or activity, patients were requested to restrain from excessive ambulation or exercise for the 12 h prior to CMR evaluations. To minimize the effect of volume overload in hemodialysis patients, MR studies were performed on the day of routine dialysis treatment within 4 h after completion of the dialysis session. Peritoneal dialysis patients had CMR on the day of their routine monthly visit.
Short-axis cine images were acquired with retrospective ECG-gating using a segmented steady-state free precession technique after localized shimming and/or frequency adjustment. Participants were asked to hold their breath as long as possible; for those who could not adequately breath-hold, images were acquired during free breathing with multiple signal averaging. For Ecc calculation, tagged cine images were acquired in the short axis at the midventricular level using an ECG-triggered segmented fast-gradient-echo sequence. Grid-tag spacing=8 mm field of view=(30–32) × (25–26) cm2, slice thickness=6 mm, flip angle=20°, TE/TR=3 ms/4.2 ms, number of segments=7 to 9; temporal resolution was 30 ms (effective frame rate ~33 frames/s). Tagged images were analyzed using the HARmonic Phase (HARP, Diagnosoft, Palo Alto, CA, USA) technique . LVM was measured and normalized to body surface area (BSA) for the LVM index (g/m2 BSA). Outcome measures included LVM index, LVH, EF, and midventricular maximum absolute circumferential strain (|Ecc|). The presence of LVH was based on CMR LVM index percentiles recently described by Sarikouch et al. . Abnormal Ecc was defined as a value <16% . The intra- and interobserver variability for peak systolic circumferential strain (Ecc) in our center is <1% .
Phosphorus-31 cardiac MRS of the septum and apex was performed in the cardiac muscle using a method previously validated . A linearly polarized transmit and receive 31P coil with a diameter of 14 cm was used. The standard phosphorus spectroscopy sequence, provided by the manufacturer with image-selected in vivo spectroscopy (ISIS) volume selection was used. The participants were positioned supine with the coil directly over the precordium, with the center of the coil at the isocenter of the magnet. Fine adjustment of center frequency (F0) was performed if the automatic F0 determination was not correct in order to ensure the correct voxel position. Shimming was performed on a volume of interest (VOI) covering the entire heart. The VOI was placed <9 cm from the center of the radiofrequency (RF) coil to limit chemical shift displacement to <10%, as previously determined . Data acquisition was done at end diastole by inserting a trigger delay. The 3D voxel of acquisition was planned to include most of the septum and apex of the heart. Voxel=80 ml (40×50×40 mm), TR=10,000 ms, number of samples=512, number of averages=136, total scan time=23 min. Spectra were analyzed and quantified on jMRUI software using AMARES, a time-domain-fitting program . Postprocessing was performed with a 15 Hz Gaussian line broadening and Fourier transformation. Phase correction was performed with PCr peak as the reference peak. Concentrations of PCr, ATP, and 2,3-diphosphoglycerate (2,3-DPG) were calculated as the area under the peaks. PCr/ATP ratio was used as an outcome variable. PCr/ATP ratio was determined after correcting the ATP peak for blood contamination.
For T2 measurements, spin-echo images of the LV in the short-axis plane were acquired using a black blood dual spin-echo method. Imaging parameters were: slice thickness=5 mm, in-plane resolution=1.4×1.4 mm, echo train length=5, echo times TE1=6 ms, TE2=34 ms. Data acquisition was done during late diastole by inserting an individual-dependent delay. The full width of half maximum of T2 distribution and mean T2 values were calculated, as described by Wansapura et al. . Outcome measurements of myocardial function from CMR and 31P MRS included: (1) Ecc, measure of myocardial strain; (2) PCr/ATP ratio, measure of myocardial energy metabolism; and (3) T2 relaxation time and FWHM, measures of myocardial structure.
Values are presented as mean ± standard deviation (SD) or median and interquartile range (IQR). A two-sample t test or Mann–Whitney rank sum test were used to compare means ± SD of continuous variables. Categorical variables were compared using the Fisher's exact test. One way analysis of variance (ANOVA) was used to compare multiple groups. The associations between variables were assessed by Spearman correlation analysis. The SAS 9.1 statistical package was used in the analysis.
Patient characteristics are presented in (Table 1). The most common kidney disease etiology in this cohort was glomerular disease: six (60%) in dialysis and six (60%) in transplant patients. Congenital anomalies/dysplasia was seen in two (20%) dialysis and three (30%) transplant patients. The remaining diagnoses were cystic disease and cortical necrosis. There were six hemodialysis and four peritoneal dialysis patients. Median time on maintenance dialysis was 9 (range 2–42) months. Three patients had previously failed kidney transplant. Four hemodialysis patients had fistulas and two had permanent atrial catheters. The mean Kt/V was 1.32±0.17 (range 1.1–2.3) for hemodialysis and 1.62±0.5 (range 1.0–2.2) for peritoneal dialysis patients. Nine patients retained their first transplant and one had second transplant; seven were treated with maintenance dialysis (median time 12.4, range 0–29 months) prior to transplant. Median time posttransplant was 21 (range 1–168) months and median duration of renal replacement therapy (dialysis + transplant) was 30 (range 2–180) months. Mean glomerular filtration rate (GFR) was 70±31 (range 37–115) ml/min/1.73 m2. Three patients had GFR >90 ml/min/1.73 m2, three had CKD stage 2 (GFR 60–89 ml/min/1.73 m2), and four had CKD stage 3 (GFR: 30–59 ml/min/1.73 m2). Immunomodulatory therapy at the time of the study included steroids (n=9), tacrolimus (n=7), cyclosporine (n=1), Rapamune (n=4), mycophenolate mofetil (n=9), and azathioprine (n=1). There was no significant difference in age, weight, height, or body mass index (BMI) among dialysis and transplant groups. A higher prevalence of anemia was observed in the dialysis group compared with the transplant group. Nineteen patients were taking blood pressure medication (mean 2.0±1.1), including angiotensin-converting enzyme inhibitors (ACEI) (n=13), calcium channel blockers (n=11), and β-blockers (n=9). Dialysis patients had more frequent uncontrolled hypertension, but the difference did not reach statistical significance (p=0.10). The majority of patients had dyslipidemia, with the most common abnormalities being low high-density lipoprotein (HDL) cholesterol (62%) and hypertriglyceridemia (48%).
Patients with CKD had significantly lower mean PCr/ATP ratio (1.25±0.73) than controls (2.18±1.01), p=0.02, Fig. 1. No significant difference in the mean PCr/ATP ratio was found between dialysis (0.99±0.48) and transplant (1.34±0.78) groups (p=0.35). No significant association between PCr/ATP ratio and LVM index was found.
Mean T2 was higher in the patient (65.1±13.2 ms) versus the control (58.9 ±8.6 ms) group (p=0.056,) although it did not reach statistical significance. CKD patients had broader mean FWHM (33.2±9.2 ms) than controls (27.9±5.6 ms; p=0.01) (Fig. 2). The T2 values were positively correlated with LVM index (r=0.46, p=0.04). No significant difference was found between dialysis and transplant patients for either T2 (p=0.35) or FWHM (p=0.90).
Age (17.9±3.1 years), height (1.67±0.8 m), weight (68.9±13.5 kg), and BMI (24.6±5.4 kg/m2) of healthy controls were not significantly different from dialysis and transplant patients (Table 1, all p values >0.10). Children and young adults with CKD had significantly higher heart rate, cardiac index, and LVM index compared with controls (Table 2). Four (40%) dialysis and three (30%) transplant patients had LVH (LVM index >97th percentile for age and sex ). There was no significant difference in LV end-diastolic volume (LVEDV) and EF between patient groups and controls. However, Ecc was significantly lower in dialysis patients versus controls and transplant patients. Nine (45%) patients had abnormal Ecc (<16%): 6/10 dialysis patients and 3/10 transplant recipients. Of seven patients with LVH, five also had abnormally low Ecc. The Ecc was inversely correlated with LVM index (r=−0.47, p=0.04).
This is the first study using CMR and MR spectroscopy to characterize early markers of cardiac dysfunction in children and young adults on maintenance dialysis and after kidney transplantation. This preliminary report demonstrates new evidence that occult cardiac dysfunction, decreased energy metabolism, and abnormal myocardial microcomposition are already present in these patients. These abnormalities were detected despite uniformly normal EF. Myocardial energy metabolism, which plays a fundamental role in the pathogenesis of cardiac disease, can be studied noninvasively by 31P MRS . High-energy phosphate metabolism derangement results in depletion of PCr and can be measured by 31P MRS as the PCr/ATP ratio. This ratio is decreased in patients with chronic heart failure, ischemic heart disease, dilated cardiomyopathy, secondary myopathies, and muscular dystrophy . Using cardiac 31P MRS, Ogimoto et al.  performed a cross-sectional study of 14 adult patients (mean age 49.5±11.7 years) on peritoneal dialysis and eight healthy volunteers. They found the PCr/ATP ratio was significantly lower in the patient group, although all patients showed normal systolic function by echocardiography. We found similar abnormalities in much younger patients (mean age 17.1±2.3 years), including patients with successful kidney transplant. One of the possible causes of decreased PCr/ATP ratio in CKD patients is an imbalance between increased energy demand of the heart and diminished energy supply in the hypertrophied heart . Many of our patients had increased LVM that might have limited their myocardial energy supply. Concurrently, higher heart rate and cardiac output might have contributed to an increased energy demand.
In our study, patients had increased T2 relaxation time and increased T2 heterogeneity, as quantified by the T2 distribution FWHM. These markers of myocardial structure are likely abnormal due to presence of myocardial water (edema) and/or increased fat content . In principle, T2 alone cannot distinguish myocardial lipid accumulation from increased water in the myocardium. Increased fat deposition in the heart should be considered, as it is known that dyslipidemia is very frequent in CKD. In fact, nearly all of our patients had an abnormal lipid profile. Proton spectroscopy, a proven noninvasive technique for in vivo quantification of intramyocellular lipids , was not performed in our study. Future studies using the technique may provide insight into myocardial lipid alterations in CKD patients. Alternatively, prolonged T2 heterogeneity due to increased water could be a marker of inflammation. This is plausible, as advanced CKD is known to be an inflammatory state, even in pediatric patients . Myocardial fibrosis could be another reason for increased T2 distribution. However, we were prevented from using gadolinium contrast, the gold standard to detect fibrosis/decreased perfusion.
Another important finding in this study was abnormally low myocardial strain in the presence of normal EF. Nearly half of the patients had strain values less than the normal cutoff. Strain analysis of MRI tissue tagging is a proven method for imaging myocardial function [31, 32]. The traditional metrics of regional and global ventricular function defined by endocardial excursion are not sensitive to early systolic dysfunction detection. Standard planimetry metrics are preload dependent. This is especially important in dialysis patients because changes in extracellular volume during and between dialysis treatments are significant. In contrast, myocardial strain is a dimensionless parameter of myocardial deformation, which has shown to be less preload sensitive.
The cause of the abnormal structure (prolonged T2) and decreased systolic function (low Ecc) may be myocyte hypertrophy. This hypothesis is supported by the results of this study that show significant association between lower Ecc and higher T2, with increased LVM index. These results are concerning, as early cardiac dysfunction was found in young patients, a population without pre-existing symptomatic cardiac disease and other comorbid conditions. This is especially concerning, as cardiac disease is the leading cause of death in patients initiated on renal replacement therapy during childhood [33–35]. The reason for increased CVD mortality in these young patients is poorly understood. Pediatric patients with CKD rarely demonstrate symptomatic atherosclerosis, the main cardiovascular morbidity and cause of death in older patients with CKD. In contrast, cardiac arrest is the major cardiovascular cause of death in children on renal replacement therapy, especially in those on dialysis. It is thought that arrhythmias are the terminal cause of cardiac arrest in these patients , but their origin is not known. We do not know whether changes in myocardial structure, function, and metabolism and their association with increased LVM found in our study contribute to an increased rate of sudden cardiac death in young patients with CKD. Nonetheless, these results are worrisome given that the presence of LVH, systolic dysfunction, and inflammation independently predict sudden cardiac death in older adults with end-stage renal disease [37, 38]. The increased T2 heterogeneity, abnormal Ecc, and low PCr/ATP ratio found in this cohort of transplant patients suggest increased risk for CVD. The majority of transplant recipients in this study were on maintenance dialysis prior to transplant. Thus, it is conceivable that cardiac abnormalities in these patients developed during the dialysis stage. We do not know whether these changes can be reversed posttransplant. If they are irreversible, this would justify a broader use of pre-emptive transplants in young patients. The majority of these patients also had CKD stage 2–3. How much this mild-to-moderate kidney dysfunction contributed to heart abnormalities is unknown. Early cardiac abnormalities were more evident in dialysis patients than in transplant recipients despite a similar level of LVH. This suggests that in addition to cardiac hypertrophy, other mechanisms might be involved in the development of cardiac dysfunction in chronically dialyzed children. The cross-sectional design and small sample size did not allow us to answer the above questions.
In this preliminary report, we found no statistical difference in the studied outcomes between hemodialysis and peritoneal dialysis groups (results not shown). However, these results should be taken with caution, as only four peritoneal dialysis and six hemodialysis patients participated. Another limitation of this study is the significant number of patients with glomerular causes of CKD known to have a higher prevalence of cardiovascular risk factors, including hypertension, dyslipidemia, and inflammation. This is in contrast to a typical CKD cause distribution in pediatric patients, with congenital abnormalities being the main cause. Therefore, our results may not be generalizable. Nonetheless, the results suggest that patients suffering from glomerular diseases may be at especially high risk for CVD progression.
In conclusion, our observations suggest that cardiac MR and MRS could play a potentially useful role in the early detection and monitoring of cardiac dysfunction in this vulnerable population. The goal of future investigations will be to determine whether the presence of subclinical cardiac abnormalities denotes a group of patients who are at higher risk of developing clinically apparent CVD and whether regression of abnormal LV geometry leads to restoration of normal cardiac function.
This study was partially presented at 2011 Pediatric Academic Society/ Society for Pediatric Research Meeting, Denver, CO, USA
Funding sources This study was funded by the research grant DK090070 from the National Institute of Diabetes and Digestive and Kidney Diseases and USPHS Grant #UL1 RR026314 from the National Center for Research Resources, NIH (M.M.M)
Disclosure The authors do not have any information to disclose.