Abstract
Biomarker use in the care of intensive care unit (ICU) patients with congenital heart disease (CHD) is a promising field, currently centered around two classes of biomarkers: (1) cardiac function/failure markers, including natriuretic peptides (brain natriuretic peptide [BNP]), suppressor of tumorigenicity 2 (ST2), and galectin-3 (Gal3) and (2) cardiac injury markers, including troponins. As cardiac function/failure markers, postoperative BNP levels correlate with adverse outcome but may not indicate severity across all CHD subtypes or ages, especially neonates or patients with Fontan palliation. BNP trends are useful for risk stratification in pulmonary hypertension. Based on accruing adult evidence and new pediatric evidence, biomarkers of cardiac failure and fibrosis, soluble ST2 and Gal3, may soon be incorporated into CHD care. As cardiac injury markers, troponins have a limited role but remain useful in resolving the mechanism of postoperative dysfunction after coronary reimplantation and discriminating active myocarditis from dilated cardiomyopathy. Data on biomarkers of noncardiac organ system dysfunction, especially acute kidney injury and neurologic injury, also continue to accrue in pediatric CHD patients. There are no established pediatric guidelines at present for guiding any therapy solely based on biomarkers. We recommend incorporating pretest probability before ordering; establishing a baseline biomarker level; evaluating trends in individual patients; and accounting for interassay variability, developmental regulation of levels, and potential confounders such as renal failure. In the future, using biomarkers as surrogate end points in clinical trials, markers of therapeutic response, and/or a means of avoiding invasive procedures could revolutionize pediatric cardiac and ICU care.
Key Words
congenital heart disease, pediatric cardiology biomarkers, BNP, troponin, ST2, Gal3, AKI
Pediatric congenital heart disease (CHD), as a heterogeneous group of rare individual diseases, with myriad anatomic variations and complicated medical and surgical management, is a prime target for biomarker discovery and applicability in perioperative patients in the intensive care unit (ICU). It is worthwhile for those involved in critical care management of patients to be familiar with common and promising biomarkers for CHD. This chapter will review the definition of a biomarker; describe challenges to discovery and validation of biomarkers in pediatric medicine; current biomarker use in CHD in the ICU; specific ICU situations, including nonsurgical and surgical care of patients with CHD and acquired pediatric heart disease in the ICU; and conclude with a literature review of biomarkers in critical care of patients with CHD.
The term biomarkers has a broad definition as “characteristics that are objectively measured and evaluated as indicators of normal biological processes, pathogenic processes or pharmacologic responses to therapeutic intervention.” For the purposes of this chapter we will use the term biomarker to refer to a protein and/or small molecule measured in a body fluid. As methods of discovery have become increasingly sophisticated, novel biomarkers are appearing with seemingly endless potential for clinical applicability. However, validation of biomarkers for everyday use in CHD is difficult. Barriers to development and validation of biomarkers include the difficulty in development of reliable, reproducible, clinically available assays; normative values in pediatrics, and developmental regulation. Biomarker discovery and validation, with the ultimate goal of creation of a high-quality assay that meets US regulatory approval, is a complex endeavor, with only the end product visible to the clinical practitioner ordering the test at the bedside. From procuring pediatric patient samples and winnowing down long lists of biologically feasible proteins to finally developing and validating reliable bedside assays on large cohorts, the process can take years ( Fig. 38.1 )
Despite the complexities of validation, biomarkers enticingly offer enormous potential for screening, diagnosis, prognosis, and therapeutic monitoring. Currently the vast majority of research and practical experience on biomarkers in CHD is centered on two classes of biomarkers: (1) cardiac function/failure and (2) cardiac injury.
Cardiac Function/Failure Biomarkers
Cardiac function/failure biomarkers fall into two broad categories: (1) biomarkers of atrial/ventricular stretch (natriuretic peptides [NPs]) and (2) indicators of myocardial fibrosis and function, suppressor of tumorigenicity 2 (ST2), and galectin-3 (Gal3). Each of these biomarkers is discussed in detail in the following sections with emphasis placed on use in the ICU.
Natriuretic Peptides
The three natriuretic peptides (NPs) include atrial natriuretic peptide (ANP), brain natriuretic peptide (BNP), and natriuretic peptide C. BNP, although also contained in the brain, is predominantly secreted by cardiac ventricular myocytes as a means of physiologic adaptation to wall stress or volume overload. The NPs are almost exclusively cardiac expressed. They each contain a common 17–amino acid ring that is essential for binding to their cognate guanylyl cyclase receptors. NPs stimulate natriuresis and diuresis by attenuating the renin-angiotensin-aldosterone axis and relaxing pulmonary vasculature by elevation of cyclic guanine monophosphate ( Fig. 38.2 ) BNP is the prototype for NPs, is the only NP studied in children, and will be the focus in the next section.
Cardiac Failure Biomarker Category 1: Natriuretic Peptides
Brain Natriuretic Peptide Structure.
BNP is the biologically active cleavage product of prohormone proBNP, and NTproBNP is the biologically inactive N-terminal ( Fig. 38.3 ). Plasma levels differ according to half-life; BNP has a half-life of approximately 22 minutes, whereas NTproBNP has a half-life of 1 to 2 hours, accounting for higher plasma levels of NTproBNP. BNP has important characteristics as a biomarker of ventricular stress and dilatation, including clinical availability of point-of-care plasma assays, stability over freeze-thaw cycles, and existing normative data for both healthy children and those with structurally normal but failing hearts, CHD, and pulmonary hypertension (PH). However, BNP is developmentally regulated, is affected by renal failure, shows wide variation between and within CHD subsets, and is affected by age, gender, and pubertal stage. The most notable developmental regulation is the log scale difference in NP levels in neonates during the physiologic transition within the first week of birth, attributed to maturation of the kidney, decreasing pulmonary vascular resistance, and acute increase in ventricular afterload after birth. There is a generalized trend of decline after the first month; adult levels are reached around 6 years of age.
Normative Pediatric Values.
Assays for NTproBNP and BNP are US Food and Drug Administration (FDA) approved and available in most clinical laboratories. For NTproBNP, Albers et al. established normative values based on review of combined data from four studies, encompassing 690 subjects from birth to 18 years of age, establishing 95th and 97.5th percentiles for healthy children. Normal values of BNP have also been established by meta-analysis of 195 healthy infants, children, and adolescents admitted for minor procedures, from birth to 17.6 years of age. In all subjects older than 2 weeks, plasma BNP concentration was less than 32.7 pg/mL. Of note, routine measurement of BNP versus NTproBNP varies by center, and even within centers because of vendor variability, further compounding the difficulty in comparison among small CHD subsets. BNP and NTproBNP are NOT interchangeable but do have excellent correlation between log BNP and log NTproBNP in adults and in children with CHD and PH. For the purposes of this chapter, BNP and NTproBNP will be referred to collectively as BNP, except where different applications or plasma levels are important.
As a broad overview of BNP in the ICU and in CHD, there are both surgical and nonsurgical applications, both of which will be discussed in more detail in this chapter, with synopsis of the existing literature. In the realm of nonsurgical care, BNP can be used in the diagnostic workup of dyspnea from respiratory versus cardiac causes, cardiac failure in congenital and acquired heart disease, and PH crisis in nonsurgical patients with pulmonary arterial hypertension (PAH). BNP is the only biomarker warranting mention in the most recent pediatric heart failure guidelines. Regarding cardiothoracic surgical care, there are preoperative, perioperative, and postoperative applications of BNP, including as a prognostic marker for outcome and major postoperative events, including low cardiac output state and postoperative PH crisis, as discussed in the following sections.
Pattern of Natriuretic Peptide Release Before, During, and After Cardiac Surgery
Preoperative Levels.
In their 2014 systematic review of the clinical utility of BNP in pediatric cardiac surgery, Afshani et al. investigated the preoperative correlation of BNP with CHD severity and early postoperative outcome, including 20 peer-reviewed studies that evaluated immediate postoperative BNP levels. In the preoperative period they found that BNP levels were associated with cardiac failure. Preoperative BNP levels were higher in those being treated for cardiac failure than in those not being treated for cardiac failure. In univentricular hearts, BNP and NTproBNP predicted clinical cardiac failure by Ross score. After data amalgamation the authors determined the area under the curve (AUC) for receiver operating characteristic (ROC) analysis of 83% (95% confidence interval [CI], 71%-95%) for the prediction of clinical cardiac failure, with an NTproBNP screening cut point of 6.75 pg/mL (sensitivity 94%) and diagnostic cut point of 86.8 pg/mL (specificity 94%); Youden index was 25.1 pg/mL. Although their systematic review showed that BNP levels predicted cardiac failure and were associated with adverse outcomes, including death, low cardiac output syndrome, length of stay, and duration of mechanical ventilation, postoperative levels were more predictive of early poor outcome (mortality <180 days after surgery). Of note, the authors’ attempted meta-analysis was unsuccessful due to difficulty in comparing results using different assays, small individual sample sizes, and heterogeneous outcome measures; ultimately, they did not seek to provide generalized cut points for diagnostic or prognostic classification in CHD.
Perioperative and Postoperative Levels.
BNP levels peak postoperatively between 6 and 24 hours. In the first few hours after computed tomography (CT) surgery, NP levels are lower than baseline, attributed to breakdown without resynthesis, removal during ultrafiltration, and decreased filling of the heart. There is a second peak in BNP levels reported around postoperative day 5.
In a prospective study of BNP levels in CHD in 221 patients undergoing CT surgery, Niedner et al. evaluated BNP levels in patients with CHD undergoing both cardiac surgery and noncardiac surgery (“surgical stress” control arm). Patients were seen at a single center from 2003 to 2005, with 103 normative controls, 13 “surgical stress” controls, and 106 CHD repairs, ranging in age from 30 weeks’ gestational age through 22 years. The purpose of the study was to identify circumstances in CHD of relative BNP deficit. They found progressively lower BNP expression in staged univentricular repair, speculated to be “perhaps from isolated cavopulmonary failure or impaired perioperative expression of natriuretic peptide” with the implication that BNP level may not be indicative of CHD severity in all types of single-ventricle physiology. The authors speculated that, especially in patients with Fontan palliation, BNP insufficiency may exist with physiologic implications. Alternatively, unloading of the systemic ventricle has also been proposed as a mechanism for BNP levels similar to those of controls. In the “surgical stress” group there was no significant difference in preoperative and postoperative BNP level, confirming that the physiologic stress of surgery alone is not a confounder. They also noted a difference in baseline preoperative BNP levels between two groups of patients, based on age: (1) neonatal patients with CHD who were less than 44 weeks’ gestational age (adjusted) and (2) nonneonatal patients with CHD who were more than 4 weeks of age plus older children. Median BNP levels in neonatal versus nonneonatal patients with CHD were 27 and 7 pg/mL, respectively. Median preoperative and postoperative BNP levels in neonatal CHD were 2370 versus 2140 pg/mL. Median preoperative and postoperative levels within the nonneonatal group were significantly different (22 versus 41 pg/mL with median change 19 pg/mL, P < .001). As mentioned earlier, postoperative BNP levels were more predictive of early poor outcome (<180 days) than preoperative or perioperative levels. However, due to small sample sizes, heterogenous outcome measures, and weak discriminatory power of AUC of ROC, ultimately cut points for diagnostic and prognostic classification are not provided in the literature.
In essence, important trends in BNP levels in patients with CHD during and after CT surgery include the following:
- •
Clear demarcation in higher BNP levels in neonatal versus nonneonatal patients with CHD
- •
Preoperative BNP levels correlated to clinical severity of heart failure and length of therapy
- •
Postoperative levels correlated to early (<180 days) adverse outcome, but
- •
Progressively lower levels in staged univentricular repair subsequent to the Norwood procedure, regardless of clinical acuity, raise the concern that BNP level may not be indicative of CHD severity in all patients with single-ventricle physiology
- •
No significant increase in levels during noncardiac surgery (“surgical stress” control group)
- •
Wide variation between and within CHD subsets; due to small size, lesion-specific subsets lacked power for statistical analysis
In summary, although BNP levels appear to be valuable as a clinical adjunct marker preoperatively and postoperatively, at this time they are used only in conjunction and support of clinical measures of severity and echocardiographic parameters, without validated algorithms for changing clinical management. Further, for individual patients, levels should be followed for trends; random levels are meaningless at best and misleading at worst.
Brain Natriuretic Peptide in Pulmonary Hypertension.
PH is an increasingly frequent cause for admission to the ICU, and PH crisis is a feared complication of postoperative management for patients with both CHD and structurally normal hearts with severe pulmonary disease. BNP is the best-studied PH biomarker in both pediatric and adult populations. However, for all the reasons described earlier, BNP has limitations, especially in the pediatric population, as have most biomarkers whose discovery has been from large-volume adult studies with results extrapolated to pediatric patients. Despite lack of a standardized catheterization laboratory protocol for pediatric PH, cardiac catheterization remains the gold standard for diagnosis, prognosis, and therapeutic monitoring in PH. In adults with PH, BNP has been correlated with disease progression as a surrogate for invasive hemodynamics, including pulmonary vascular resistance and pulmonary artery pressure, with a negative correlation with cardiac index. In pediatric PH, however, although elevated/rising BNP level is recognized as a means for high-risk versus low-risk stratification, conclusive correlation with invasive hemodynamics has not been demonstrated. Three studies in pediatric patients that attempt to correlate BNP with invasive hemodynamics had mixed results. Bernus et al. found no correlation with echocardiographic or hemodynamic data, whereas Lammers et al. found optimal correlation between pulmonary vascular resistance index (PVRI) and BNP when PVRI is lowest and most stable. Finally, although Takatsuki et al. demonstrated correlation of NTproBNP with invasive hemodynamics, the low magnitude of the correlations led the authors to recommend against using levels to replace clinical parameters. Finally, as noted earlier, BNP and NTproBNP are well correlated but not interchangeable, and different centers tend to collect either one or the other, rarely both, with more centers now collecting NTproBNP than in prior years. This is relevant given the assertion by Takatsuki et al. that although BNP was better correlated with real-time hemodynamics in the cardiac catheterization laboratory, NTproBNP is preferable for longitudinal monitoring with longer half-life and stable levels.
Robust correlation with functional parameters and surrogates for clinical end points in pediatric PH would be highly valuable, but at best the data are conflicting. Lammers et al. found in a retrospective cohort of 50 children with PH that BNP levels did not correlate with 6-minute walk distance (6MWD) testing, whereas Van Albada et al., in a cohort of 29 pediatric patients with idiopathic PAH (IPAH) and associated pulmonary arterial hypertension (APAH), found that NTproBNP levels correlated with functional class and 6MWD, and initiation of treatment resulted in decreased NTproBNP levels and increased 6MWD.
Further, in adult IPAH, BNP levels greater than or equal to 180 ng/mL predict worse survival outcomes, but attempts in pediatric PH to identify a similar clinical cut point have proven more difficult, presumably due to smaller sample sizes, longer time until outcomes of interest, and poor sensitivity. Lammers et al., using AUC from ROC analysis of the UK Pulmonary Hypertension Service for Children, found that a BNP level of greater than 130 pg/mL predicted death or need for transplant with a sensitivity of 57.1% and specificity of 83.3%. Bernus et al., when attempting to extrapolate the adult value of 180 ng/mL, had only 10 patients with BNP level of greater than 180 ng/mL. Finally, Van Albada et al. attempted to evaluate an adult PH cutoff value of NTproBNP level greater than 1400 pg/mL to identify patients with poor long-term prognosis (53% sensitivity and 88% specificity), but only 6 pediatric patients had values of greater than 1400 pg/mL. Using this cut point of greater than 1400 pg/mL, there was 83% mortality within 2 years; using a cut point of NTproBNP level greater than 1664 pg/mL gave 100% and 94% sensitivity and specificity, respectively, for mortality.
In summary, in PH:
- •
Trends of BNP levels for individual patients appear more useful than a random absolute value at any given time.
- •
A threshold BNP value of BNP 130 pg/mL (NTproBNP >1664 pg/mL) is likely valuable for risk stratification.
- •
Despite mixed results and low-magnitude correlation with functional parameters and invasive hemodynamics, a rising/elevated value within the framework of developmental regulation implies higher risk of adverse outcome.
Currently there is no validated use of biomarkers specifically for PAH associated with CHD (APAH-CHD). It will be especially difficult to generalize and/or create clinically useful cut points for APAH-CHD, given heterogeneity, small subsets of specific lesions, and the difference between elevation of pulmonary pressure in biventricular hearts and that in single-ventricle physiology. In a recent expert consensus statement on PH in children with CHD by the European Pediatric Pulmonary Vascular Disease Network, based on the 5th World Symposium on Pulmonary Hypertension in Nice in 2013 and the Pediatric Taskforce of the Pulmonary Vascular Research Institute (Panama, 2011), there is no mention of biomarker use to guide therapy. Thus extrapolation of the results discussed earlier to patients with CHD should be done carefully and with the understanding that no expert consensus or validated studies exist.
Cardiac Failure Biomarker Category 2: ST2 and Gal3
As described previously, the two general classes of cardiac failure biomarkers are those of ventricular stretch (NPs), and those of cardiac function and fibrosis, namely, ST2 and Gal3. Both ST2 and Gal3 are gaining widespread acceptance as adult biomarkers of myocardial function, fibrosis, and remodeling based on a growing body of literature and warrant discussion for anticipated future pediatric cardiology applications, particularly in heart failure and orthotopic heart transplant. Moreover, increased utility has been demonstrated as an adjunct to BNP in heart failure in adults because both were included in the 2013 American College of Cardiology/American Heart Association guidelines as risk stratification biomarkers for acute and chronic heart failure in adults (class IIb recommendation). However, increased concentrations of both biomarkers have been noted with concomitant noncardiac pathophysiologies such as chronic obstructive pulmonary disease and pneumonia, likely reflecting an inflammatory, and potentially confounding, milieu in patients with multiorgan system comorbidities.
Soluble ST2 Structure and Function.
ST2, a member of the interleukin-1 receptor family, is a receptor for interleukin-33 (IL-33). Two isoforms of ST2 are involved in cardiac signaling and pathogenesis: soluble ST2 (sST2) and a cell membrane–bound isoform (ST2L). Binding of the membrane-bound ST2 with its ligand IL-33 results in cardioprotective signaling. However, soluble ST2 functions as a scavenger receptor and competes for IL-33 binding with the membrane bound ST2. When sST2 levels are high, IL-33 binding to the cardiac membrane bound ST2 is reduced; in the absence of IL-33/ST2L cardioprotective signaling, there is resultant cardiac cellular death, tissue fibrosis, and reduced cardiac function. sST2 has indeed been shown to be a marker of cardiac cellular death, tissue fibrosis, and reduced cardiac function in adult heart failure and transplant patients, as well as being predictive of hospitalization and mortality.
Normative Pediatric Values.
Normative median and 95 percentile ST2 values in children without heart failure were found to be similar to those of normal adults. Although not statistically significant, higher levels were noted in males versus females when they were 15 years of age or older.
Pediatric Use.
ST2 shows promise as a potential biomarker of orthotopic heart transplant rejection in pediatric heart transplant recipients and hopefully in the overarching goal of eventual biopsy-free detection of rejection and therapeutic monitoring of antirejection medications. In a recent simultaneous biopsy and serum-based assessment of ST2 in both heart and small bowel transplant, Mathews et al. showed that sST2 level was elevated in rejection and quiescent during rejection-free periods. The efficacy of this potential biomarker in two types of organ transplant supported claims that elevated sST2 level was a reflection of alloimmunity rather than just cardiac injury. Immunostaining of endomyocardial biopsy specimens with marked increase in ST2 level during rejection, coupled with moderate discrimination by AUC of ROC analysis of serum at the time of rejection episodes, supports the circulating biomarker as a reflection of proximal graft rejection. Further, serum levels of ST2 returned to rejection-free levels with effective treatment. Finally, given that sST2 does not require cardiomyocyte damage for secretion, such as other markers of cardiac injury, it is further proposed as a marker of early rejection.
Galactin-3 Structure and Function.
Gal3 is a carbohydrate-binding protein released by activated cardiac macrophages, resulting in the induction of cardiac fibroblasts, and is found to be upregulated in decompensated heart failure. This complex interplay, with progressive accumulation of myocardial collagen and alteration of the myocardial extracellular matrix, is hypothesized to link the inflammatory milieu of myocardial injury to fibrosis. Gal3 is also implicated in apoptosis, which is speculated to be involved in the transition from compensated to decompensated heart failure. Gal3 is an example of a biomarker that likely also plays an active role in the pathogenesis of heart failure and may have a future role as a therapeutic target or agent.
Normative Pediatric Values.
In children without heart failure, Gal3 levels are similar to those in previous studies in normal healthy adults. Levels varied according to which FDA-approved assay was used; as previously mentioned as a common weakness in biomarker validation, interassay variability impedes establishment of pediatric normative values and is a source of discrepancy among biomarkers. Gal3 is not correlated with gender but does show a positive correlation with age.
Pediatric Use.
As in adults, Gal3 has potential applicability in risk stratification for pediatric heart failure across causes, including CHD, but as of yet there is no current validation for its use in clinical management. Two studies (by Kotby et al. and Mohammed et al. ) have evaluated serum Gal3 levels in children, including patients with preserved and reduced ejection fraction in dilated cardiomyopathy, CHD, and rheumatic heart disease. Both studies found a statistically significant difference in Gal3 levels between those with heart failure, whether with preserved or reduced ejection fraction, and control patients. Analysis of AUC of ROC showed a cut-off value for differentiating patients with heart failure versus controls of greater than 3 ng/mL with sensitivity of 100%, specificity of 97.78%, positive predictive value of 97.8%, negative predictive value of 100%, and diagnostic accuracy of 100%. Significant differences were found in both studies between Gal3 levels and severity of heart failure by the Ross classification. Kotby et al. also found a statistically significant increase in Gal3 levels of those not receiving spironolactone, whereas in adult patients receiving Aldactone, Gal3 levels were decreased. Thus an increase in Gal3 levels is speculated to be due to Gal3-mediated myocardial fibrosis. Given the prevalence of Aldactone as an adjunct in heart failure in both pediatric and adult CHF patients, this is an unfortunate confounder, but it also indicates that elevated Gal3 levels may identify patients who would reap the most benefit from the antifibrotic effect of spironolactone.
Biomarker of Cardiac Injury: Troponin
Due to unparalleled cardiac specificity, troponins serve as acute cardiac myocyte necrosis markers and are currently the best circulating measure of acute cardiac injury. As a broad overview, use of troponin in the ICU in CHD centers around identification of cardiac ischemia, postpericardiotomy syndrome, cardiac trauma, and myocarditis.
Troponin: Structure/Function.
Cardiac troponin is the calcium-dependent regulator of the contractile apparatus of cardiac muscle, often thought of the “switch” for cardiac muscle contraction and relaxation. Three subunits constitute troponin: troponin C (calcium-binding subunit [cTnC]), troponin T (thin filament that attaches to tropomyosin [cTnT]) and troponin I (inhibits actin-myosin interactions [cTnI]). The troponins are uniquely cardiac specific and are therefore specific cardiac necrosis biomarkers. Both cTnT and cTnI have been studied in pediatric and adult cardiac patients, and both are released with cardiac injury, with cTnI being unaffected by renal failure and thus the preferred of the two.
Normative Pediatric Values.
Assays for cTnI and cTnT are FDA approved and available in most clinical laboratories. Interassay variability exists, but generally the clinical upper limit of normal is considered the 99th percentile of the normal reference population. Based on the 2002 American College of Cardiology and European Society of Cardiology guidelines, this is approximately 3 standard deviations from the mean of a normal adult population, to maximize sensitivity while minimizing false-positives. Typically the upper reference limit cutoff of 0.04 ng/mL is based on apparently healthy adults and used to risk stratify myocardial infarction in adults at The Johns Hopkins Hospital. Normative values exist in pediatrics, with the highest concentrations immediately after birth, peaking on the third day of life, and decreasing toward adult values by the end of the first year of life. Reported values in the literature tend to be more helpful to describing trends of developmental regulation, given the variety of assays used and increasing sensitivity of the assay over the decades since discovery. In an attempt to establish normal pediatric values, Hirsch et al. evaluated two groups of children, the first of which included ambulatory pediatric patients with no apparent cardiac disease ( n = 120) and patients in stable condition with known congenital or acquired cardiac abnormalities ( n = 96), whereas the second group included 65 ICU patients with a variety of acute illnesses, with the results that cTnI levels were generally not elevated in group A. As will be further described later in the chapter, troponin levels are higher in infants than older children.
Cardiac Ischemia in the Intensive Care Unit.
In both surgical and nonsurgical patients, as well as those with and without structural heart disease, cTnI and cTnT are the preferred biomarkers of cardiac injury in both adults and children, with improved sensitivity and specificity for cardiac injury over creatine kinase (CK) and creatine kinase–myocardial band (CK-MB) assays. cTnT is a sensitive and specific marker of myocardial injury, given that it is not expressed by skeletal muscle, either during development, as a result of skeletal muscle injury stimulus, or following noncardiac surgery. In pediatric CHD, cTnI and cTnT are sensitive and specific for myocardial ischemia/infarction, although cTnI is often preferred because levels are not affected by renal failure. In pediatric cardiology, cTnI is less likely to be elevated from myocardial ischemia due to coronary artery disease, given the much lower prevalence of coronary artery disease in pediatric patients. As discussed previously, in patients in stable condition with known congenital or acquired cardiac abnormalities, troponin levels were within the same range as those in their apparently healthy, ambulatory pediatric peers. In a recent study by Harris and Gossett of 24 patients with elevated troponin levels, excluding recent cardiac surgery, “significant” CHD, neonates in the neonatal intensive care unit, or patients on extracorporeal membrane oxygenation (ECMO), only 3 had coronary-related diagnoses (anomalous origin of left coronary artery from pulmonary artery [ALCAPA] and Kawasaki disease), whereas the majority (17/24, or 71%) had myocarditis or cardiomyopathy. Although left heart catheterization was completed in nearly half the cases (10/24), in no case was the diagnosis made/changed, confirming the current standard of care among pediatric cardiologists that left heart catheterization and coronary angiography is reserved for a highly selective group, avoiding routine application of adult “door-to-balloon time” protocols. Currently troponin is used as a clinical adjunct in the diagnosis of pediatric myocarditis; however, normal levels do not rule out the presence of myocarditis. Further, in a retrospective review of patients presenting to the emergency department with chest pain over 7 years by Brown et al., only 48% of those with increased troponin levels were attributed to primary cardiac disease, although the yield increased to 87% when combined with abnormal ECG. The median troponin level of those with cardiac cause of chest pain differed from those with a noncardiac cause: 8.5 ng/mL (interquartile range 4.6-18.5) versus 0.34 ng/mL (interquartile range 0.12-1.3) ( P = .02), respectively. Two other important conclusions were drawn from this study: (1) less likelihood of noncardiac diagnoses in patients if troponin level is greater than or equal to 2 ng/mL and (2) no correlation between level of troponin elevation and morbidity and mortality. Both of these conclusions were supported by a recent study evaluating the contribution of diagnostic modalities toward the final diagnosis after an elevated troponin level was demonstrated, as well as the contribution of the troponin level at various times during diagnosis and treatment to the final diagnosis and ventricular function. This study found that troponin values within each subgroup did not distinguish between cardiac causes, with a wide range of values, among which levels in myocarditis patients were not the highest. Troponin absolute value and progression were also not helpful in differentiating between preserved versus depressed ventricular function.
Although these studies were done in patients with no known history of CHD, given the findings of the study by Hirsch et al. of comparable baseline troponin levels in normative versus stable cardiac disease, they can likely be extrapolated to CHD patients.
Thus, although the differential of cTnI elevation in both congenital and acquired heart disease is broad, including coronary artery aneurysm, injury, anomalous origin of the coronary arteries, anthracycline-induced cardiac toxicity, myocardial surgical injury, myocarditis, postpericardiotomy syndrome, and trauma, cTnI elevation in children is fundamentally different from that in adults, although adult cutoff values for normalcy are applied. As with NTproBNP, knowledge of baseline values and/or trends over time are often more clinically relevant than a random value in time, especially given the demonstration, by increasingly sensitivity assays, that there is development/accrual of morbidity from structural and acquired heart disease, resulting in slowly rising troponin levels over time. This inherently decreases the utility of solitary cutoff values, even in adults. Further, experts in the field have theorized that an early release of troponin may signify reversible injury, whereas sustained release may be associated with progressive cell death. Values trended over time are therefore more informative, and an important future goal of risk stratification will be recognizing a pattern of troponin release that may distinguish acute disease/injury from chronic elevation.
Pattern of Troponin Release During Cardiac Surgery.
Postoperative troponin levels can have value to explain poor cardiac function, especially in cases that require coronary reimplantation. Troponin release in CHD following cardiothoracic surgery and cardiopulmonary bypass (CPB) has been studied, with the following five themes: (1) cTnI and cTnT are of equal sensitivity and specificity for myocardial injury; levels behave similarly in the early postoperative phase (up to 28 hours). cTnT has been described as peaking 30 minutes after termination of CPB with steady decline to baseline 4 to 5 days postoperatively ; (2) troponin I is preferable to cTnT as a biomarker of cardiac ischemia because it is unaffected by renal failure; (3) whereas infant myocardium is more resistant to hypoxia, congenitally abnormal myocardium is more sensitive to cardioplegic arrest, with higher postoperative peak troponin levels in infants ; (4) higher postoperative values are associated with longer myocardial ischemia and extent of myocardial damage ; and (5) threshold levels for troponin as a marker of adverse outcome have been difficult to establish in children.
Similar to BNP, troponin levels have been used in adults as a marker of poor outcome after CPB, with values above threshold levels being associated with severe cardiac event and/or death. Similar thresholds have been proposed in pediatric cardiothoracic surgery for both cTnI and cTnT, with description of high troponin I release (especially ≥ 100 mcg/L) being associated with post-CPB mortality. However, as discussed previously, as assays become increasingly sensitive, thresholds established using older assays become obsolete, and possible pitfalls of attempting to distinguish thresholds, especially across heterogeneous patterns of cardiac injury, has the potential to mislead. Further, in a contemporary study of troponin values in infants and children with CHD undergoing CPB, Gupta-Malhotra et al. showed that, with the caveat of small sample size, for the 4 patients with cTnI values above this threshold (3 infants, 1 child), none had a severe cardiac event or death. Further, infants appear to have higher postoperative values of troponin, making thresholds difficult across age ranges. Finally, differing results have been found in preoperative and postoperative troponin levels in cyanotic versus acyanotic patients, further complicating establishing a threshold value across the heterogeneous spectrum of CHD. As described earlier, the adult cutoff value of cTnI less than 0.04 ng/mL is extrapolated to pediatric patients but does not account for developmental regulation, individual variance in cardiac injury, or presence of cyanosis.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
