Echocardiogram screening in pediatric dialysis and transplantation

Hypertension is the most common CV risk factor present in children with chronic KF, with uncontrolled hypertension evident in 51% of children after 1 year of dialysis [10] and in 25% 5 years post-transplant [11]. Hypertension causes endothelial injury and arterial stiffness leading to an increase in peripheral vascular resistance and pressure load on the left ventricle [6]. Over time, this causes myocardial injury, ventricular hypertrophy, and remodelling. Hypertension is also a pro-atherogenic state [6].

Dyslipidaemia, another pro-atherogenic state, has been reported in 45% of the pediatric transplant population [12]. An increase in atherogenic lipoproteins is seen, with inhibition of antioxidant pathways. Pro-inflammatory effects contribute to endothelial dysfunction and vascular stiffness [13]. Disturbances in body mass index (BMI) are common in KF, with coronary artery calcification more prevalent with obesity and malnutrition [14, 15]. Metabolic syndrome is increasingly prevalent post-transplantation and this can contribute to vascular inflammation, arterial stiffness, and atheroma [16].

Chronic KF is a complex state of persistent low-grade inflammation [17]. Pro inflammatory cytokines are elevated, with C-reactive protein (CRP) associated with coronary calcification in children [18], and interleukin-6 (IL-6) and CRP associated with CV mortality in adults [19]. Dialysis is an independent exacerbator of inflammation by mechanisms including peritoneal dialysate and hemodialysis membrane bio-incompatibility, and catheter-associated infections [17]. Fluid overload has also recently been associated with markers of inflammation [20].

Elevated serum calcium, phosphorus, and parathyroid hormone (PTH) are all independently linked to increased risk of coronary artery calcification [18] and carotid intima media thickness (CIMT) [21] in children with chronic KF. Increased ventricular mass can be linked to cellular growth triggered by FGF23, as a result of impaired phosphate and vitamin D metabolism [22]. In addition, there are challenges with achieving an ideal balance of treatment, with overtreatment with activated vitamin D and calcium containing phosphate binders associated with CIMT [23] and coronary artery calcification [24].

Anaemia is a frequent complication of dialysis and reduced GFR and has been associated with LVH and diastolic dysfunction in several pediatric studies [21, 25, 26]. Cardiovascular complications of anaemia are believed to be due to a chronic increase in cardiac output, poor tissue perfusion, and reduced oxygen delivery to the myocardium [27].

Dialysis is perhaps the most significant risk factor for CVD. Pre-emptive transplantation and reduced cumulative time on dialysis have been associated with improved survival [3, 28]. A multitude of inter-related factors are at play including exaggerated risks of traditional and uremic/inflammatory factors, and of course fluid overload. Risk is potentially most significant for the hemodialysis population, with extremes of volume overload and intra-dialytic hypotension associated with mortality in adults [29]. Temporary global or regional reduction in systolic myocardial function, called myocardial stunning, is seen frequently in children on conventional hemodialysis, and can be related to intradialytic hypotension with reduced coronary perfusion and changes in volume loading [30]. Repeated myocardial stunning may cause chronic injury, contributing in some circumstances to chronic systolic dysfunction [31].

Dialysis patients are at risk of pressure and volume overloaded cardiovascular states, both of which have been associated with increasing LV mass, hypertrophy, and functional changes including impaired ventricular relaxation (diastolic dysfunction) and ventricular contraction (systolic dysfunction). There is significant interdependence of hypertension and volume overload, although control of volume state may be the most important factor in reducing LV mass [32]. The geometry of LVH may give a clue as to the predominant risk state, with volume overload and ventricular dilatation leading to a more eccentric hypertrophy, and pressure overloaded systems characterised by a concentric hypertrophy [33]. Arterial stiffness identified by increased aortic pulse wave velocity (PWV) can contribute to this pressure overload. It is often described as a marker of vessel ageing, which is seen in children with KF [34]. With a lifetime of kidney replacement therapy ahead of them, these changes are significant.

Trans-thoracic echocardiography is the most widely used diagnostic modality available for structural and functional cardiac assessment. Conventional modes of echocardiography use two-dimensional (2D) assessments of structure and motion to assess myocardial wall thickness, wall mass, chamber size, and systolic function. Accurate assessment requires gating with an electrocardiogram to determine timing of systole and diastole. Integration of Doppler measures allows evaluation of pressure, velocity, valvular function, and diastolic/systolic function (including tissue Doppler imaging TDI). Newer methods such as strain imaging and three-dimensional (3D) echocardiography may provide more accurate assessments and early detection of sub-clinical systolic dysfunction [35]. Overall echocardiogram limitations include inter- and intra-observer variation, indirect derivation of some measurements, and system and resource considerations. Specific to KF, volume state dependence of some measures needs to be taken into consideration. A summary of echo modalities, clinical utility, and limitations is discussed below and further detail of commonly reported parameters in chronic KF publications is provided in Table 1.

Wall thickness, chamber size, and functional assessment should form part of each echocardiogram assessment and there are published recommendations to guide this [39], as well as to standardise reporting [53].

Given the potential hemodynamic changes in chronic KF, echocardiogram standardisation can be a challenge [54]. Ideally, routine assessment of patients on dialysis should occur at rest, on an interdialytic day when close to or at ideal/target weight, and at target hemoglobin concentration. Ideal weight, however, can be difficult to assess and echocardiography may also be used to help define this. Inferior vena cava (IVC) parameters along with clinical correlation have been shown to accurately predict fluid status in dialysis patients with adult studies showing improved cardiac mechanics and left ventricular mass (LVM) with dry weights adjusted according to IVC diameter [55]. Additional data can be derived from echocardiograms performed during dialysis or in the immediate post-dialysis period [30, 56] and a number of methods incorporating time have been used to reduce the impact of volume status on echocardiogram measures. An example is the myocardial performance index (MPI), or Tei index. It is assessed using Doppler imaging and is defined as the sum of the isovolumic contraction and relaxation times divided by the ejection time [38]. It is considered a reliable parameter for global LV function assessment but is mainly used in a research capacity in children. Other echocardiogram confounders include age, gender, race, and body composition. To account for body composition, LVM is commonly indexed to body surface area (BSA) or height, and presented as the left ventricular mass index (LVMI). To compare with healthy reference populations, values in pediatric echocardiography have been recommended to be presented as Z-scores [39]. Given body composition issues in chronic KF, presentation of height or BSA-based Z-scores for LVM may be most accurate [57]. The most relevant reference intervals should be used which are best representative of one’s cohort, and in accordance with one’s vendor’s equipment. There may, however, still be challenges in finding the most appropriate reference for the individual given the substantial variability in the healthy population, particularly for LVM.

An increasing array of echocardiogram changes in chronic KF are now described. In many of these, associations with cardiac events and mortality have been reported in adult populations. Longer outcome studies in pediatrics are required. Detailed definitions of more commonly described markers in clinical and research use are provided in Table 1.

There are countless observational studies of prevalence and risk of echocardiogram changes in pediatric chronic KF. Many incorporate control populations to demonstrate significantly worse LVM, diastolic and systolic function [21, 41, 58‐61]. Each publication must be interrogated closely. Demographic factors such as duration of dialysis can significantly impact findings, and definitions of abnormal values vary. Table 2 presents key studies evaluating children with chronic KF, aiming to highlight the prevalence of echocardiogram abnormalities, multivariable analysis-derived risk factors, and follow-up of echocardiogram changes over time.

LVM/LVH is the most commonly studied echocardiogram parameter, with some studies also assessing LVM influence on function, demonstrating a correlation with worse diastolic [21, 59], and systolic [66] function. Hypertension has been consistently identified as a key risk factor for LVH in dialysis [8, 25, 58, 63, 64, 70] and transplant [70, 71] populations, with overweight/obese BMI also associated with LVH in dialysis [25] and transplant populations [61, 71]. LVM risk appears to increase from CKD to dialysis in children, although longitudinal studies are not available. In both peritoneal dialysis (PD) [65] and hemodialysis (HD) [8, 25] volume-related factors, anaemia [25, 58, 65] and hyperparathyroidism [21] are key risks. Where these are carefully managed, LVM and LV geometry can improve [57]. With transplantation, some regression in LVM can also be seen [72], although hypertension and metabolic risks can lead to persistent changes, or deterioration over time [71, 73]. A recent study has correlated LVM in children on dialysis with pulse wave velocity (PWV), a marker of vascular stiffness [74]. A direct causal relationship between PWV and LVM/LVH has not been established, and the correlation may simply reflect that hypertension is a common risk factor. CIMT is also associated with LVH in PD patients [65], again suggesting the presence of a common risk state. In this publication, inflammation was postulated as important given that both LVH and CIMT were significantly associated with CRP [65].

Global LV systolic function as measured by conventional echo-based LVEF/FS is mostly preserved in childhood chronic KF [35, 75]. During HD sessions, however, regional myocardial systolic dysfunction assessed by SF% (stunning) is a common finding [30]. In adults stunning has been demonstrated to progress to fixed regional systolic dysfunction and global reduction in LVEF by 12 months [31] and is associated with 12-month mortality [76]. Impaired GLS measured by STE has been demonstrated in childhood chronic KF in the absence of conventional echocardiography measures of systolic dysfunction [56]. In adult dialysis patients, GLS is associated with CV mortality independent of age and conventionally measured EF [43]. The HD population appears to be most at risk, especially if assessed during or immediately following a HD session [35, 47, 56, 77] and if there are more significant falls in systolic blood pressure or higher ultrafiltration volumes [56, 77]. Segmental differences in LV strain can also be seen at this time [47] including where measured by the asynchrony index (Table 1) [77]. This is a marker of LV mechanical dyssynchrony which is where there are regional differences in timing of ventricular contraction and relaxation. The asynchrony index is associated with ventricular arrythmias and sudden cardiac death in adult dialysis patients independent of LVEF [78]. A recent study using 2D STE post-hemodialysis has shown impaired LVEF derived by STE versus normal LVEF by conventional echocardiogram in children, providing further evidence of the sensitivity of STE [47]. Strain by STE or TDI has been shown in some pediatric transplant cohorts to be significantly worse than controls [35, 59], and has been associated with hypertension, metabolic syndrome, obesity, dyslipidaemia, lower EGFR, and past HD (versus PD or pre-emptive transplant) [61]. Longitudinal patient assessment has demonstrated that GLS can improve with transplantation [75], although it remains significantly worse versus controls [61]. Mitsnefes et al. have demonstrated evidence of LV hypercontractility in dialysis [8] and transplant patients [79]. In dialysis, the contractile reserve in exercise was reduced compared with controls, whereas in transplant patients the reserve was maintained. It was postulated that this could suggest an adaptive response perhaps mediated by sympathetic overactivity, and that over the long-term, this may be disadvantageous. To date, however, associated long-term outcomes have not been assessed to our knowledge. Kim et al. identified in the post-transplant cohort poorer LV performance, defined by the Tei index, and worse GLS by TDI, in the setting of increased FS and EF, and that this was associated with a longer time on dialysis prior to transplant [59]. Again, no longer-term follow-up data is available.

LV diastolic function can be impaired in both dialysis and transplantation [21, 41, 59, 60, 80]. Most studies evaluate this as a continuous parameter using the ratios E/A, E′/A′, or E/E′, although some define specific criteria for diastolic dysfunction [41, 60, 64, 66]. TDI measures of diastolic function may be more sensitive than PWD [41]. Duration of dialysis has been associated with markers of LV stiffness [80], and diastolic dysfunction (DD) may persist post-transplant [59, 60, 75]. DD has been associated with LVH/LVMI in HD [66] and transplant [67, 79]; however, this is not a consistent finding [41, 73], and could reflect how LVMI or DD was defined/measured, or other risk factors for DD in the studied population. Markers of volume overload [65], and hyperparathyroidism [21] have been associated with worse diastolic function. Risk post-transplant may be independent of hypertension [41, 73]. In adults, DD measured by E/E′ ratio has been associated with mortality [81]. In HD, left atrial strain (LAS) measures may be an even more sensitive measure of DD and cardiac events [82]. LAS has recently been demonstrated to worsen during HD in children [56].

Right heart changes are infrequently reported in childhood chronic KF, although these are still an important part of standard echocardiogram assessment. Right heart geometry has been reported to change in adolescents on HD, especially in the presence of an arterio-venous fistula [83], with increased right atrial and right ventricular (RV) free wall thickness and reduced RV volumes reported. This may be due to an increase in venous return.

Ascending aortic dilatation appears to be somewhat of a novel finding for children (versus adults) with chronic KF [68, 84, 85], and it is also seen in childhood CKD [86]. Poor nutrition may be an important risk factor [86]. Pre-emptive transplantation may be protective, and there appears to be an association with post-transplant hypertension [85]. We were unable to find publications with aortic measurements reported in adult chronic KF populations.

Guidelines that incorporate routine echocardiography assessment of chronic kidney failure patients are listed in Table 3. These have mainly focused on pre-transplantation screening of adult patients to determine suitability, as cardiac mortality in the first few months post-transplant can be high [87]. Identification of ischaemic heart disease (IHD) via non-invasive tests including stress echocardiograms is the predominant focus, although this is controversial, especially for asymptomatic patients. The American Society of Transplantation guidelines incorporate LVEF into their risk stratification for progressing to non-invasive IHD screening in transplant candidates, with routine echocardiogram screening suggested for those who are identified at risk based on clinical screening, X-ray and ECG [88]. The NKK-KDOQI guidelines are the only guidelines that recommend more regular echocardiography in dialysis [89]. In addition, these are the only guidelines that reference children, in whom a resting echocardiogram is recommended in the first 3 months following dialysis initiation [89]. No guidelines address routine echocardiogram screening in transplant recipients.

While prevalence of echocardiogram abnormalities is significant across childhood dialysis and transplant populations, current pediatric recommendations for echocardiogram screening in chronic KF are limited. A key screening test principle is that screening followed by an intervention should reduce clinical events. Currently, there are no trials assessing whether echocardiogram triggered interventions improve CV events. Physiologic principles, longitudinal studies, and multivariable analyses, including adult studies and those assessing echocardiogram outcomes, do, however, suggest that some interventions could be of benefit. For example, improved volume state and blood pressure in children is associated with better LVM in PD and HD [25, 63]. In adults, more frequent HD has been associated with reduced myocardial stunning [92], and in longitudinal studies, LVH regression with improved systolic function [93]. Shroff et al. demonstrated an increase in LVMI over 12 months for children on HD, with no significant change seen for patients on hemodiafiltration (HDF) [25], and Fadel identified an improvement in diastolic and systolic function for children converted to online-HDF with associated decrease in CRP [69]. In adults, some studies have shown improved survival with HDF [94]. Low hemoglobin has been associated with LVMI/LVH in childhood dialysis [25, 58], and in adults, correction of anaemia has been associated with regression of LVH [95]. Hemodialysate cooling was associated with less myocardial stunning in a randomised crossover trial in adults [96], and in a cross-sectional study of adults, stunning was rare in PD versus HD [97].

One could argue that many echocardiogram abnormalities are associated with volume overload or hypertension, and that a screening echocardiogram should not be required to identify these states. Regular echocardiogram screening to identify subclinical and potentially reversible end-organ damage could be of benefit, however, to prompt an earlier change in management and a more complete evaluation of patient CV risk and therapy options. This could be individualised with screening further extended or increased in frequency according to local echocardiogram access or baseline suspected risk. For example, HD patients could be screened during dialysis or immediately post-dialysis to identify stunning or strain. Given that stunning has been associated in adults with irreversible systolic dysfunction and mortality by 12 months [31], and there are indicators that strain is associated with myocardial fibrosis, any accessible intervention that has some level of evidence and is without harm should be considered. For example, change in modality to PD may be associated with less cumulative risk of stunning/strain, and more frequent HD or hemodialysate cooling may reduce stunning. Transplantation may also reduce these risks, with some evidence that impaired strain can be reversed [61]. Post-transplant, CV screening recommendations include metabolic and BP screening [98, 99], with ambulatory BP screening recommended due to the higher risk of masked hypertension in this population [99]. Of note, based on the American Association of Pediatrics (AAP) hypertension guidelines, any child with hypertension and kidney disease should have a yearly echocardiogram [99].

Minimum recommended standards for pediatric echocardiography have been published by the American Society of Echocardiography [39, 53]. These standards include all the important assessments from a chronic KF perspective, including measures of the left ventricle and aorta, an assessment of systolic function, and Doppler-based diastolic function. Given that longitudinal chronic KF studies identify evolving echocardiogram changes within 12 months [25, 61, 69, 70], a minimum of annual frequency of echocardiogram screening seems appropriate but can be adjusted according to findings and risk factors.