Dennis M. Fischer, M.D., Editor
The concept that fetal and neonatal myocardia are functionally immature developed several decades ago, based on experiments showing that fetal and neonatal myocardia develop less tension per unit cross-sectional area than adult myocardium. [1–4] In addition, neonatal hearts responded less than adult hearts to most pharmacologic and physiologic stimuli, [1,3,5–7] and neonatal hearts require higher extracellular Ca sup 2+ concentrations to optimize contractility. [1,2,8–11] Similarly, the response of immature human myocardium to anesthetics differs from that of adults, and neonates appear to be more sensitive to the depressant effects of volatile anesthetics. [12–14] These characteristics can relate to differences in cellular structure and physiology.
Although the advent of fetal echocardiography has made possible qualitative and quantitative measures of immature human myocardium, knowledge of specific human myocardial development remains incomplete. Complicating extrapolation of animal research to humans is the fact that late fetal cardiac development differs among species. Myocytes from rabbits, [15,16] rats, [17] dogs, [18] and cats [19] are relatively immature at birth, whereas myocytes from sheep [20] and guinea pigs [21] are relatively mature. Many programmed changes in cardiac development and control occur (Table 1), and the reasons for many of these are as yet unknown.
Recent investigations into the mechanisms and regulation of cardiac function have revealed increasingly complex levels of control as our understanding has progressed from the morphologic to the physiologic and more recently to the molecular biologic level. This review will present developmental aspects of cardiac function and relate it to anesthetic effects on the immature heart. Readers interested in additional information on subcellular control mechanisms are referred to recent reviews by Lynch. [22,23]
Whole Organ Function
Systolic Function
Force generation by isolated muscle strips in a variety of nonhuman species, standardized for cross-sectional area, increases during development. [1,3,7,24,25] In humans, force development in the intact fetal heart is significantly less than in the adult, and the decreased fractional shortening observed by fetal echocardiography is similar to the decreased shortening in isolated strips of nonhuman fetal myocardium. [26,27] The fetal heart has limited increases in output in response to changes in preload [28–30] and afterload. [19,28] Instead, fetal combined ventricular output increases significantly with increases in heart rate. [31,32] At birth, left ventricular output increases two- to threefold, [24,33,34] a result of alterations in preload, afterload, contractility, and heart rate. [34,35]
Studies of isolated fetal muscle strips have shown that fetal myocardium cannot generate the same contractile force as adult myocardium throughout the entire range of the length-tension curve. [7,36] Optimal end-diastolic sarcomere length, perhaps the ultimate measure of preload, is age-independent. This was originally ascribed to the relative deficiency of contractile elements, but as described in later sections, immaturity of a variety of subcellular systems and limited sympathetic modulation of cardiac function may also contribute. After birth, the neonatal heart has a limited ability to increase cardiac output in response to alterations in preload or afterload, [24,37–39] probably as a result of already near maximal beta-adrenergic stimulation. Contractile reserve increases as myocytes mature, and baseline beta-adrenergic tone decreases. [24,34,37]
These findings led to the claim that the immature heart is extremely heart rate-dependent, i.e., it is capable of modulating cardiac output via alterations in heart rate and is relatively insensitive to changes in intravascular volume, i.e., preload. [40] This is an oversimplification. Whether preload (the Frank-Starling mechanism) or heart rate is the primary regulator of cardiac output in the fetus and newborn remains controversial. Clearly, both mechanisms are functional.
Studies in fetal and immature experimental animals have produced conflicting results. Some investigators have reported that heart rate, rather than preload, is the primary modulator of cardiac output in the immature animal, as increases in preload increase stroke volume and cardiac output minimally. [29,41] Several studies have failed to demonstrate a significant increase in right or left ventricular output in response to a volume load. [19,28,30] This may result from the fetal heart operating near the peak of its Starling curve (to be discussed). [29] There is a rate-dependent increase in contractility in immature myocardium in vitro, [42] although the mechanism for this is unknown.
Others, in contrast, have reported that the Frank-Starling mechanism has primacy even in the immature heart and that heart rate changes have little, if any, effect on output. [31,43] A Doppler study in human fetuses has suggested that the Frank-Starling relationship is the prime regulator of cardiac output in the human fetus, as in adult humans. [44] The resting end-diastolic pressure in fetal ventricles is about 3–5 mmHg. The plateau of the Frank-Starling relationship in the fetal ventricle appears to be at a lower pressure than in the mature heart (Figure 1). [29] Difficulties in studying fetal cardiac hemodynamics include the involvement of the compliant utero-placental bed and the possibility that volume infusions in early studies had a secondary effect in increasing afterload by filling an already maximally dilated bed. When left ventricular afterload is controlled, increasing left atrial pressure beyond 10 mmHg increases stroke volume. [45] The neonatal animal, unlike the fetal animal, is capable of a significant increase in cardiac output in response to a volume load, although this increase is more pronounced after the first few weeks of life. [38] Clinicians recognize that anesthetized neonates routinely increase blood pressure with acute volume infusions.
The response of neonatal myocardium to changes in afterload is difficult to determine because of the interrelationship of afterload and preload. At very low levels of afterload in the intact heart (as opposed to isolated muscle strips), preload is not maintained, and cardiac output decreases; thus in these studies preload must be rigorously controlled. Fetal and neonatal myocardium are more sensitive to changes in afterload than adult myocardium, and the fetal right ventricle appears to be extremely so. [19,28]
Doppler studies of the human fetal heart confirm previous experimental studies in the fetal lamb, namely that the right ventricle predominates as early as the late first trimester. [26,27,46] Similarly ultrasonographic and Doppler studies of the human fetus show that right ventricular output is 60% of the combined ventricular output in late-term human fetuses, [47] similar to early work in fetal lambs. Ventricular output near term (approximately 450 ml [center dot] kg sup -1 [center dot] min sup -1) is similar in fetal humans and fetal lambs. [41]
Diastolic Function
Diastolic function, particularly as it relates to anesthesia, has recently been reviewed. [48,49] The diastolic characteristics of the heart relate not only to sarcomeric diastolic function, but also to passive stretch properties of the cytoskeleton, intercellular matrix, ventricular interaction, and pericardium. Compliance of isolated myocardium and intact ventricles increases with maturation. [3,7,25,36,50] Both -dP/dt (the change of pressure with time during diastole), a measure of early relaxation, and tau (the time constant of isovolumic relaxation), a measure of later diastolic filling, are decreased in neonatal hearts compared with adult hearts. [51] Measurements of tau may be more relevant at the rapid heart rates of young animals because relaxation from the previous systole will typically not have any effect on end-diastolic pressure at heart rates of less than 150–160 beat/min. [52]
The findings of decreased diastolic compliance in isolated nonhuman fetal myocardium have been confirmed in Doppler studies of human fetuses. [53,54] In addition to intrinsic properties of immature myocytes and connective tissue, the lung-chest wall and the pericardium limit right and left ventricular filling in the fetus in vivo. [55]
Relaxation of the myocardial cell requires diffusion of Ca2+ from troponin C. If sarcoplasmic Ca2+ is not rapidly reduced by the sarcoplasmic reticulum (SR), uptake, then relaxation, will be limited. Thus the poorly developed SR in immature myocardium (to be discussed) may have an important role in the impaired lusitropy (myocardial relaxation) of fetal hearts. The large surface-to-volume ratio in immature cells may permit a greater role of sarcolemmal removal of Ca2+ than in more mature cells. Most of the Ca2+, at least in mature cells, is removed by the Na sup +-Ca2+ exchanger, [56] and neonatal cells have more exchanger protein, as described below.
Ryanodine, an inhibitor of SR function, inhibits relaxation. Relaxation is inhibited to a greater extent in mature, than in immature, hearts, again consistent with the functional immaturity of the SR in young hearts. [50] This suggests that alternative mechanisms for the removal of Ca2+ are relatively more important in immature myocardium. The affinity of troponin C for Ca2+ would also be a factor in determining relaxation and may relate to the developmental changes in troponin I and troponin C during maturation. It has been suggested that mitochondrial Ca sup 2+ uptake is greater in the newborn rabbit than in the fetus and adult. [1]
Both ventricles interact via the common ventricular septum. Increases in right ventricular pressure or volume can inhibit left ventricular diastolic filling. A decrease in right ventricular preload or an increase in right ventricular afterload may increase left ventricular filling, affecting cardiac output. [35,57]
O sub 2 Consumption, Metabolism, and Coronary Flow
Myocardial metabolism differs between immature and mature myocardium. In vivo, myocardial O2consumption in newborn sheep and rabbits is higher than in adults and is associated with greater myocardial blood flow. [51,58,59] Higher O2consumption in newborns is commensurate with increased cardiac work (estimated as the rate-pressure product). Matching of coronary flow to O2consumption is similar between age groups, as evidenced by similar O2extraction ratios.
Developmental changes in mitochondrial ultrastructure parallel changes in myocardial substrate use. The most important source of energy to the adult myocardium is long-chain fatty acids. The enzyme responsible for the transfer of fatty acids into the mitochondrion is carnitine palmitoyl CoA transferase. Activity of this enzyme is decreased in immature myocytes. As a result, the primary energy substrates in immature myocardium are lactate and carbohydrate, [58,60] thus explaining, at least in part, mitochondrial-dependent age-related increases in aerobic metabolism.
Electrophysiology
Developmental changes are also evident in pacemaker and conduction system cells. Neonates have a higher intrinsic rate of sinus node automaticity, but the basis of this has not been delineated. Sinus node action potential duration also increases with age. In one study, maximum diastolic potential was significantly less negative (closer to threshold) in infant rabbits than in adults, [61] but in another study, this was not the case. [62] Threshold potential did not differ between infants and adults. [61,62] Neonates have a shorter atrioventricular (AV) conduction time that increases with maturation as a result of increases in intraatrial and His-ventricle conduction times. [63,64] AV nodal conduction time, however, is not different between neonates and adults. Although infant and adult AV nodes are morphologically distinct, their action potential (AP) characteristics are similar. [63]
Histology
During the fetal and early newborn periods, the heart grows by hyperplasia (increasing cell number). After this period and throughout the rest of its life, the heart can increase mass in response to demand for increased pumped volume or increased pressure only by myocardial cell hypertrophy. [65,66]
Myocardial cells (cardiocytes) change their gross appearance and their ultrastructural components during development. Cardiocytes in the very immature heart are initially smooth and rounded, changing to rod-like in the maturing cell. With continued development, the cell shape becomes irregular and develops intercalated discs, with their associated gap junctions. In humans, gap junction genes are expressed by 15 weeks of gestation. [67] Relative to total heart weight, there is more connective tissue and less contractile tissue in immature myocardium. Sarcomeres comprise 30% of fetal heart in contrast to 60% of adult heart. The surface area-to-mass ratio is greater in immature cells [68] as is the water-to-collagen content.
Whereas mature cardiocytes contain parallel myofibrils within each cell, the immature myocytes have a smaller fraction of myofibrils, [1,2,7,16,20] which are less organized and may not be arranged parallel to the long axis of the cell. [15,20,65] With development, myofibrils align with the axis of the cell but lie at the periphery of the cell surrounding centrally located nuclei and mitochondria. Later, myofibrils distribute throughout the cell, as in the adult configuration.
Known maturational changes in the SR are thought to be a primary explanation for differences in contractility and in Ca2+ responsiveness of neonatal versus adult myocardium. The sarcoplasmic reticulum of the immature myocardial cell is markedly underdeveloped as compared with mature myocardial cells (Figure 2). [15,16,69,70]
T-tubules are invaginations of the sarcolemma of ventricular, but not atrial, cardiac myocytes that extend longitudinally within sarcomeres, allowing for more direct contact of the sarcolemma with subcellular structures lying at a distance from the cell surface. This allows for simultaneous transmission of the AP, with its associated ion shifts, to all parts of the cell, facilitating rapid activation of the entire cell. T-tubules are a requirement, therefore, of larger cells. [71] The state of the t-tubule system at birth is species-dependent. Species that have mature myocytes at birth, such as the guinea, pig have well-developed t-tubule systems, whereas those animals born with relatively immature myocytes, such as the rabbit, do not.
Mitochondria develop in absolute size and relative intracellular volume during myocardial cell development, [16,18,20,72] which may continue into the postnatal period. [16,20] In addition, the intramitochondrial cristae lengthen and become more compact.
An intracellular construction of microtubules and microfilaments links the contractile filaments, the t-tubules, sarcolemma, mitochondria, and nuclei. This structure, composed predominantly of the proteins titin and actinin, determines cell organization and size and allows the force of contraction to be applied to the myocyte. Developmental changes in the organization of the intracellular organelles probably reflect, and are mediated by, changes in the cytoskeleton, [73] but are largely undefined.
Subcellular Components
Studies at the subcellular level indicate that Ca2+ mechanisms that activate contraction in the immature heart (which lacks a well-developed SR) derive Ca2+ primarily from extracellular sources. The various sub-cellular mechanisms such as ion channels, exchangers, and pumps (Figure 3) are optimized to achieve this end.
Calcium Transport Mechanisms
Major Mechanisms. Although Ca2+ entering via the Ca2+ channel contributes to maintenance of the AP plateau, it has a critical role in supplying Ca2+ directly and indirectly to support excitation-contraction coupling. Because of the long-known relative immaturity of fetal SR, the sensitivity of immature myocardium to Ca2+, and of the exaggerated dependence of neonatal myocardium on extracellular Ca2+ to maintain inotropy, [8] sarcolemmal Ca2+ channels have been an area of interest in developmental cardiac electrophysiology. Although the current-voltage relationship for the major, large (L-type) Ca2+ current, ICa,L, is similar in immature and mature ventricular myocardium, [74] the kinetics (the time-dependent changes in channels allowing entry or exclusion of ions) of this channel, and its subunit composition, vary with development in the rat and rabbit, where it has been particularly studied. [75–79] This may account, at least in part, for the different pharmacologic properties of neonatal and adult Ca2+ flux. The smaller T-type Ca2+ current (ICa,T) is less prevalent in neonatal myocardium. [74]
The primary function of ICa,L in mature myocardium appears to be activation of the SR Ca2+ release channel with consequent release of large amounts of Ca2+ that have been stored in the SR, so-called Ca2+-induced Ca2+ release, [80] which in turn supports contraction. In immature myocardium, ICa,L likely contributes more directly to activator Ca2+, which probably explains the well-known sensitivity of neonatal hearts to the depressant effects of clinically used Ca2+ channel blockers. [81–83] Verapamil, for example, although efficacious in terminating supraventricular tachycardia, is contraindicated during infancy for this reason. [84,85] After cardiopulmonary bypass, neonates are more sensitive to the depressant effects of rapid infusions of platelets and fresh frozen plasma, presumably as a result of chelation of Ca2+ by the citrate anticoagulant in the plasma fraction of banked blood. Citrate has known direct Ca2+ channel depressant action, and this depressant effect can be prevented or reversed in the laboratory, and clinically to a great degree, with additional Ca2+. [86] Although ICa,L is of clear and increased importance in immature myocardium, a number of lines of evidence suggest that immature myocardium requires Ca2+ source(s) in addition to ICa,L to generate tension:
1. Isolated neonatal cardiac myocytes contract in the presence of a Ca2+ channel blocker. [87,88]
2. Force of contraction of neonatal rabbit right ventricular papillary muscles is unaltered by the Ca2+ channel agonist BAY K8644. [89]
3. Although some Ca2+ entering the cell via ICa,L may bind to the contractile apparatus, the magnitude of this Ca2+ is, in itself, inadequate to support contraction. Ca2+ entering via ICa,L is estimated to result in only approximately 10% of the Ca2+ available for tension generation in the adult myocardial cell and would result in far less than the 1.4 micro Meter required for 50% myofilament activation in neonatal myocardium. [75]
4. The contractility-voltage relationship of isolated neonatal cells is linear, rather than mirroring the parabolic shape of the ICa,L current-voltage relationship. [11]
5. Ca2+ current (ICa,L) increases throughout development [75,76,90,91]
6. Density of Ca2+ channels also increases, [74,76,90,92,93] although there are some species differences. [77](The predominant L-type Ca2+ current, ICa,L, has recently been evaluated in atrial myocytes from children 3 days to 4 yr of age. [94] The density of ICa,L is lower than that reported in adult atrial myocytes, although it does not change with age within this age range.).
The search for an alternative Ca2+ supply to immature myocardium has focused primarily on the Na sup +-Ca2+ exchanger. The Na sup +-Ca2+ exchanger acts as the primary mechanism for extruding Ca2+ from the cell after each contraction, thereby allowing for diastolic relaxation. Because three Na sup + ions are exchanged for each Ca2+ ion, the current is electrogenic and responsive to membrane potential. The driving force for this flow is the intracellular-to-extracellular Na sup + gradient. However, the exchanger is reversed at potentials around the AP peak and plateau, allowing Ca2+ entry into the cell.
Nakanishi and Jarmakani [95] and Hoerter [16] speculated that cardiac Na sup +-Ca2+ exchange activity would be greater in immature myocardium, and this has been confirmed: there is an age-dependent decrease in Na sup +-Ca2+ exchange activity during development of the chick and rabbit;[95,96] Na sup +-Ca2+ exchanger mRNA is six to eight times higher in late-term fetuses and neonates than in adults in rabbit and rat;[97] exchanger current density is also higher in neonatal rabbits, which have a poorly developed sarcoplasmic reticulum at birth, than in neonatal guinea pigs, which have a more mature sarcoplasmic reticulum present at birth. [98] Further, exchanger protein is abundantly located in the sarcolemma in neonatal and fetal rabbit myocardium. After the development of the t-tubules (related to increased Ca2+ control by the SR), the exchanger protein is found associated with the t-tubules, as it is in mature myocardium. [99] In addition, the prolonged AP duration in immature myocardial cells [100](Figure 4) would allow prolonged inward entry of Ca2+ via the Na sup +-Ca2+ exchanger. A single study in humans, however (representing two hearts at 17 and 19 weeks gestation), showed decreased expression of Na sup +-Ca2+ exchanger mRNA [101] compared with adult hearts.
Chin et al. studied contraction and intracellular [Ca2+] in isolated, single neonatal rabbit ventricular myocytes. They estimated that Ca2+ influx via the Na sup +-Ca2+ exchanger accounts for only 60% of the increase in intracellular Ca2+ with contraction (but far more than the 10–20% in adult hearts). [102] Nevertheless, it appears that reliance on Ca2+ entry via Na sup +-Ca2+ exchange represents a significant difference from the adult.
Supportive Mechanisms
A variety of other subcellular systems maintain an internal milieu conducive to maximal entry of transsarcolemal Ca2+ via the exchanger and L-type channel, and optimal effect of that calcium in excitation-contraction coupling.
Because intracellular Na sup + is exchanged for Ca2+ by the Na sup +-Ca2+ exchanger, maintaining an appropriate Na sup + gradient is essential. The Na sup +-K sup + pump (Na sup +-K sup + ATPase) uses energy from ATP to maintain an appropriate Na sup + gradient across the sarcolemma. The Na sup +-K sup + pump contains the inhibitory binding site for the cardiac glycosides such as digoxin. Different molecular isoforms of this subunit exist in rats [103,104] and humans, [105] and activity increases with maturation in rats and sheep. [96,106,107] The functional importance of these maturational changes is not known, but it is of clinical interest that children tend to require higher plasma concentrations of digoxin than do adults and tolerate these higher plasma concentrations without toxicity. [108,109] Cardiac glycosides, by inhibiting Na sup +-K sup + ATPase, increase intracellular Na sup + concentration and thus decrease Na sup +-dependent Ca sup + efflux, resulting in higher intracellular Ca2+ levels and presumably increased inotropy. There is an increased inotropic effect of these agents in the fetal and neonatal rabbit compared with the adult, [110,111] consistent with a prominent role for the Na sup +-Ca2+ exchanger in modulating inotropy in the immature myocardium.
A sarcolemmal Na sup +-H sup + exchanger uses the Na sup + gradient across the sarcolemma to transport H sup + ions out of the cell, although its exact contribution to maintaining intracellular pH is not known. Na sup +-H sup + exchange is greater in neonatal, than in adult, rabbit and rat hearts. [112,113] This may explain the greater function during acidosis [114,115] and the greater recovery from an acidotic insult of the neonatal myocardium. [113,116] In addition, increased Na sup +-H sup + exchange during acidosis may decrease the Na sup + gradient, permitting more Ca sup + accumulation.
ATP-dependent Ca2+ pumps remove Ca2+ from the cytoplasm to maintain an extremely low cytoplasmic concentration (1/10,000 that of extracellular fluid). High-affinity, low-capacity, Ca2+ pumps in the sarcolemma extrude Ca2+ from the cell, whereas a high-capacity, lower affinity, ATPase located in the membranes of internal organelles, such as sarcoplasmic reticulum, pump Ca2+ into the SR lumen.
The SR of the immature myocardial cell is functionally and morphologically underdeveloped in immature myocardial cells. [16,70,117] Nakanishi et al. showed that Ca2+ release by the SR is minimal in fetal and neonatal myocardium. [1,70] The number of Ca2+ ATPases, Ca2+ uptake, Ca2+ release, the activity of the Ca sup 2+ pump, and the efficiency of the Ca2+ pump increase during maturation in rabbits, [1,118] guinea pigs, [50] and sheep. [119–121] Consistent with the small role of SR in immature myocardium, Ca sup 2+ depletion of the SR by ryanodine has its most pronounced inhibitory effect on contraction in the adult, less in the newborn, and least in the fetus. [122,123]
In mature myocardium, the SR has junctional and free components. The junctional SR contacts the sarcolemma at the cell surface or at the t-tubules. Calsequestrin, a Ca2+ storage protein, is located within the junctional SR and binds large numbers of Ca2+ ions but with low affinity, allowing maximal storage but easy unbinding and egress with opening of SR Ca2+ release channels (ryanodine receptors). Calsequestrin activity increases with cardiac maturation. [118]
The various parts of the SR are connected by strands of free SR, which are rich in ATP-consuming Ca2+ pumps and phospholamban, a regulatory protein that is first expressed at the time of the first contractions of the embryonic heart. [124] Unphosphorylated phospholamban inhibits the SR Ca2+ pump, reducing the rate of Ca2+ accumulation within the SR. Phosphorylation of phospholamban by one of three protein kinases relieves Ca2+-ATPase inhibition and thus facilitates cell relaxation. [125] At least one of these, protein kinase C, is under developmental control. [126] Two of these kinases phosphorylate phospholamban in response to beta-adrenergic stimulation [127] and thus are sites at which catecholamines can exert a positive lusitropic effect. [127] Phospholamban is decreased in the SR from fetal sheep, [119] and only low levels of phospholamban mRNA are expressed in human neonatal atrial myocytes. [94]
Although the immature myocardium was reported to lack Ca sup +-induced Ca2+ release, [102,128] a recent study of human atrial myocytes found that Ca2+ channel-dependent Ca2+ release (i.e., SR function) appears in these cells by age 3 days (the youngest age studied);[94] however, the rate of release increased with increasing age, suggesting ongoing maturation. Given ultrastructural differences between atrial and ventricular myocytes, applicability to ventricular myocytes is unclear. Immature human ventricular myocytes have not been studied.
Contractile Proteins
A variety of developmental changes in the contractile proteins, constituting the thick and thin filaments of striated muscle, confer differing Ca2+ sensitivity with cardiac maturation. Each of these proteins can be expressed as several distinct isoforms that differ in amino acid sequence and in function. Isoform expression is determined by species [129] and cell type (skeletal vs. myocardial; atrial vs. ventricular) and is modulated by physiologic signaling (e.g., thyroid hormone or the development of diabetes [130]).
Myosin, the most abundant contractile protein, comprises approximately 60% of myofibrillar protein. It is responsible for transducing chemical energy (ATP) into mechanical energy. Each myosin is composed of two heavy chains and two pairs of light chains. The tails of the heavy chains are woven together to form the thick filament. The globular heads of the heavy chains project to form cross-bridges with actin and contain the ATPase activity that contributes to fiber shortening with contraction. The increasing tension developed by myocardium with increasing age correlates with increasing activity of myosin ATPase. [131]
Cardiac myosin can be resolved into three forms, each consisting of two heavy chains, alpha and/or beta. The V1 isoform (two alpha chains) has the highest ATPase activity. [132] The V2 isoform (an alpha and a beta chain) is intermediate, and the V3 isoform (two beta chains) has the lowest ATPase activity. Because ATPase activity is associated with maximal shortening of the myofibril, changes in isoforms have obvious physiologic sequelae. [131] In all species of fetal mammals studied, the beta myosin heavy chain predominates. [133] This is advantageous to the fetal myocardium as the beta isoform consumes less ATP to develop the same force. [134] The development of myosin expression is species-dependent. [129] In the human ventricle, the V3 isoform is the most abundant fetal form (90% at 30 weeks gestation), decreases after birth, and reaccumulates after the second month of life such that V3 is predominant in adult life. [135,136] Atrial myosin, with its briefer contraction, is primarily V1.
It appears that developmental regulation of heavy chain expression is at the transcriptional level and accounts, at least in part, for the very rapid conditioning of the left ventricle induced by pulmonary artery banding in preparation for an arterial switch operation in older neonates and infants with transposition of the great arteries. [137] In rats, systolic, or a combination of systolic and diastolic, overload results in a shift from the V1 to the V3 isoform, seen as early as 3 days after imposition of increased workload. In humans, in whom V3 normally dominates, no shift is seen, although humans lose the small amount of V1 normally present. [137–139] In human atria, which have primarily V1, hypertrophy results in a decline in V1 with a concomitant increase in V3.
Located near the head portion of the myosin molecule, the myosin light chains consist of alkali (LC1) and phosphorylatable regulatory (LC2) subunits, which regulate myosin ATPase activity, with an additional fetal ventricular alkali subunit isoform, LCemb, [135,140] identical to that in atrium and in fetal skeletal muscle. In humans, this form decreases in the postnatal period as the other LC1 type increases, becoming the only light chain form in adult ventricles. As with many such developmental molecular biologic discoveries, the physiologic implications of the shift to the adult isoform is not yet known.
The thin filament consists of a polymer of two intertwined helical strands of actin monomers. There are two actin isoforms: skeletal alpha-actin and cardiac alpha-actin. During human cardiac development, both isoforms are present in myocardium. [141] In the early embryonic heart, the noncardiac form predominates. The cardiac form is present in the early heart tube and increases gradually. The onset of rhythmic contraction coincides with disappearance of the noncardiac isoform. In the human heart, skeletal alpha-actin increases to 50% in the first decade. [141] Skeletal alpha-actin increases rapidly after imposing left ventricular outflow obstruction before declining slowly.
Tropomyosin lies in the groove between the two actin polymeric strands and, depending on its deformation by troponin (to be discussed), either permits or prevents the interaction of actin with the myosin head group. It exists as a homodimer or heterodimer of the two isoforms, alpha and beta. The alpha 2 homodimer is found in ventricular myocytes from small animals, [142] and large animals have predominantly alpha 2 isoform with smaller amounts of the beta isoform. There is an inverse relationship between intrinsic heart rate and isoform composition. The amount of beta isoform increases from 5% in fetal ventricles to 10% in adult ventricles in humans. [143] The physiologic effects of this developmental shift are not known. Troponin is a complex of three distinct, but functionally coupled, proteins that confer Ca2+ sensitivity to cross-bridge formation, each with multiple developmentally and species-related isoforms: troponin T [144,145] binds the complex to tropomyosin, with a relationship between isoform and Ca2+ sensitivity;[15,146,147] troponin C is the Ca2+ binding component (binding of Ca2+ results in dissociation of troponin I from troponin T, leading to cell contraction), whereas troponin I regulates interaction of the complex with tropomyosin in animals [148] and humans. [149,150] The isoform shift to the adult form of troponin I is complete in humans by age 9 months. [149]
There is preliminary evidence that the shift in troponin I isoforms is modulated by sympathetic innervation. [151] The cardiac, but not the skeletal isoform, is phosphorylated in response to beta-adrenergic stimulation. Phosphorylation decreases the affinity of troponin C for Ca2+, decreasing sensitivity of myofibrillar ATPase to Ca2+ and altering contractility. [152] In addition, the slow skeletal troponin I isoform in immature hearts may contribute to the relative resistance of the neonatal myofilament to deactivation at acid pH, [148,153] again perhaps explaining the greater recovery of neonatal myocardium from an acidotic insult.
Regulation of the Action Potential
Rapid depolarization of the cardiac AP is mediated by Na sup + influx through its channel, whereas decreased resting K sup + conductance, and subsequent sequential activation of a variety of other K sup + currents, modulates tile AP plateau and repolarization.
Resting membrane potential (RMP) and AP characteristics show developmental changes that are species- and tissue-dependent. [154] There are developmental changes in RMP, Vmax (rise of AP upstroke), AP amplitude, AP duration, and AP morphology. [154–157]
With maturation, there are increases in Na sup + conductance and accelerated Na sup + channel kinetics. [158] The distinct kinetics of the Na sup + channel in immature myocardium suggest there could be altered effects of the class I antiarrhythmics, which act by blocking Na sup + channels. [158]
Resting (diastolic) potential and the AP repolarization rate are regulated by a variety of K sup + channels. Many newer antiarrhythmics drugs have effects at K sup + channels. Maturational changes in IK1, the inward rectifier that controls primarily RMP, include lower channel conductance. [90,156] Resting depolarization increases with hyperkalemia, acidosis, or cardiac glycosides. However, immature cardiac cells are less affected by increased extracellular K sup + concentration than are adult cells, [156,159] possibly because of the lower conductance of IK1and perhaps a larger relative role for other ions in determining RMP in neonates. [156] The spike-and-dome morphology of the AP is absent in neonates and gradually appears over the first few months of life in humans (Figure 4) and other species. [160,161] The transient outward K sup + current, Ito, which is largely responsible for this morphology, is absent in neonates. [161] This alteration in AP morphology maintains the initial plateau of the AP, enhancing Ca2+ entry via Na sup +-Ca2+ exchange, as described previously. The current density of IKATP, the ATP-sensitive K sup + current, also increases from neonate to adult. [162] The phenomenon of ischemic preconditioning is thought to be mediated via these IKATPchannels, and increased activity of this channel has recently been implicated in the tolerance of chronically hypoxic neonatal hearts to ischemia. [163]
Autonomic Regulation
Responsiveness of the heart to sympathetic stimulation appears to be species-dependent. [3,6,7,9] In humans, a chronotropic response to epinephrine and isoproterenol begins at 5 weeks of gestation. Sympathetic adrenergic cardiac nerves can be detected as early as 9–10 weeks gestational age. [164] Postnatally, at least in sheep, the increasing contractile response to adrenergic stimulation with maturation is a result of higher resting adrenergic tone in the newborn rather than intrinsic maturation of the myocardium. [34] The increase in inotropic state reflects increased beta-adrenergic stimulation, [34,165] which can be prevented by fetal thyroidectomy near term. [166] Because thyroid hormone has been shown to increase beta-adrenergic receptor binding in immature rabbits [167] and rats, [168] thyroid hormone may mediate the increased beta-adrenergic stimulation at birth.
beta-adrenergic receptors and the adenylyl cyclase system are well developed by late fetal life (Figure 5). [93,169,170] beta -receptor density peaks at birth and decreases postnatally, with the decline mirroring maturation of cardiac sympathetic innervation, which is incomplete at birth in rabbits [171] and dogs. [172] Nevertheless, Ca2+ channels in neonatal myocardium respond less to beta-adrenergic stimulation than in the adult. [173] The ability of isoproterenol to stimulate adenylyl cyclase is decreased in the late fetus and neonate, [170,174] which may be a result of decreased coupling of the beta-adrenergic receptors and adenylyl cyclase at birth because the maximum inotropic response to the direct activator forskolin is greater than that to isoproterenol. [6,175]
There is preliminary evidence that sympathetic innervation of immature myocardial cells accelerates the kinetics of the Ca2+ transient and contraction and that sympathetic innervation may induce a shift from skeletal to cardiac troponin I. [151] The developmentally regulated cAMP-dependent protein kinase (PKA) phosphorylates a variety of molecules, including troponin I, phospholamban in the SR, and the sarcolemmal Ca2+ channel (ICa,L), [151] all serving to increase Ca2+ delivery to and from the contractile apparatus, with increases in inotropy and lusitropy.
The SA node is sensitive to beta-adrenergic stimulation in the fetus. In the rat, sensitivity decreases in the late fetus at the SA node but not until the first postnatal week in the ventricles. [176,177] This parallels development of autonomic innervation of each of these regions. [170]
Human newborns, even those who are preterm, have well-developed vagally mediated cardiac responses to hypoxemia and other stimuli. In humans, acetylcholine inhibits sinoatrial pacemaker activity in hearts as young as 3–7 weeks gestation, and parasympathetic cholinergic nerves are found in the atria as early as 8 weeks of gestation. Sensitivity to acetylcholine increases up to week nine and then remains stable until the eighteenth week. [164] However, in vivo, the earliest fetal response to maternally administered atropine occurs at 15–17 weeks with increases in the magnitude of the response as gestation continues. [164] Vagal myelination progresses throughout fetal development in humans and reaches adult levels by about 50 weeks postconceptual age. [68]
Expression of cholinergic receptors, which appear to be similar to adult receptors, is maximal at birth, remains high for several weeks, and then declines to adult levels, at least in the rat. [93,178] The number of muscarinic receptors peaks 1–2 weeks earlier than beta-adrenergic receptors.
Anesthetics and the Immature Heart
Anesthesia and Ion Currents
Volatile and a variety of intravenous anesthetics affect many of the voltage-dependent myocardial ion currents that have been studied. Most studies have been performed only in mature myocardium. No studies have directly examined anesthetic effects on sarcolemmal Ca2+ flux in immature myocardium, although halothane can inhibit ICa,L in fetal and neonatal myocardium as it does in mature myocardium (Figure 6). Baum and Klitzner [42] found that halothane and isoflurane decrease height of the plateau of the AP of newborn right ventricular papillary muscle, which is consistent with an effect on transsarcolemmal Ca2+ entry. No studies compare the dose responses of neonatal and mature myocardial ICa,L to anesthetics. A variety of anesthetics are known to not only inhibit ICa,L but also to shift activation and inactivation characteristics of this current. [179] These changes might decrease Ca sup 2+ entry via ICa,L in cells with more negative resting potential, such as immature myocardial cells.
BAY K8644, a Ca2+ channel agonist can only partially reverse or prevent depression in neonatal rabbit right ventricular papillary muscles caused by halothane or isoflurane. [89] This is consistent with the view that the volatile anesthetics inhibit myocardial function by depressing systems in addition to ICa,L even in neonatal myocardium. Baum and Wetzel [180] showed that halothane, in clinically relevant concentrations, reversibly inhibits Na sup +-Ca2+ exchange in neonatal ventricular myocytes. This provides an additional mechanism that may be responsible for the more pronounced depression by volatile anesthetics of immature myocardium with its increased reliance on Na sup +-Ca2+ exchange.
Although a variety of volatile and intravenous anesthetics have effects on the various K sup + currents, [95,181] there is no specific information regarding the interaction of anesthetics on K sup + currents in immature myocardium.
Anesthesia and Sinus Rate
Infant rabbit hearts in vitro are more resistant than adults to direct depression of sinus node pacemaker rate by halothane and isoflurane. [51] Maximal depression of heart rate is approximately 10% in infants. These findings suggest that direct effects of the agents on the sinus node are probably not responsible for bradycardia seen during clinical anesthesia. It is likely that cholinergic influences play a role in some cases of bradycardia. During induction of anesthesia with halothane and nitrous oxide in infants and children, cholinergic tone is present, as demonstrated by a dose-related increase in heart rate after atropine administration. [182]
For adult and infant rabbit hearts, sinus rate is more resistant to depression by halothane and isoflurane than are other indices of myocardial function. [51] These anesthetics decrease spontaneous pacemaker rate by decreasing the rate of diastolic depolarization and increasing AP duration. [183] In adult rabbit sinoatrial node, 1% halothane decreased the rate of diastolic depolarization, and 2% halothane further decreased this rate but also moved maximum diastolic potential (V sub m) closer to threshold potential (Vth). These opposing effects counterbalance to produce little change of sinus rate, [184] which may explain the relative resistance of heart rate to changes by this agent. Depression of IKand ICa,L may account for the effects of volatile agents on the sinus node, but it is likely that other ion channels are also involved.
Anesthesia and AV Conduction
Halothane and isoflurane prolong AV conduction time directly. [51] This effect is greater with halothane than isoflurane and greater in infants than in adults. For isoflurane, the difference between age groups is small and may be the result of slight differences in intraatrial, AV nodal, or His-ventricular conduction. For halothane, the difference is marked and likely represents significant differences in direct actions on AV nodal conduction. Because AV nodal depolarization depends on the slow inward Ca2+ current, the differential effect between age groups may be a result of an age-related sensitivity to Ca2+ channel blockade, although this is not known.
A recent in vitro study of rabbit hearts showed that at pharmacologically relevant levels, there was no effect of propofol on atrial or AV nodal conduction in neonatal hearts, although propofol prolonged AV nodal conduction in adult hearts. [185]
Anesthesia and O sub 2 Consumption and Metabolism
In vitro, halothane and isoflurane increase coronary flow in a dose-related manner in infant rabbit and fetal lamb hearts. [51,186] In the isolated heart preparation in which coronary perfusion pressure is constant, an increase in flow indicates a decrease in coronary vascular resistance and therefore direct vasodilatory actions of the agents on the coronary vessels. Changes in coronary flow are similar between newborn and adult hearts. Other in vitro studies have reported an increase in coronary flow by isoflurane and variable effects by halothane in adult animals. [187–190] Reactivity of coronary vessels to other pharmacologic and metabolic stimuli have been demonstrated in newborn animals of several species. [191–194]
Isoflurane decreases O2consumption and the O2extraction ratio, which, coupled with increased coronary flow, indicates decreased autoregulation and relative overperfusion of the myocardium, demonstrable in newborn and adult hearts. [51,188,195] Because it decreases heart rate more in adults, isoflurane decreases O2consumption and extraction more in adults than in newborns. [51] When heart rate is held constant by pacing, there are no differences in this effect between age groups.
In the neonatal lamb undergoing hypoxic stress, neither halothane [196] nor isoflurane [197] alter redistribution of blood to vital organs, including the heart. In addition, myocardial blood flow in the neonatal lamb decreases significantly at 1 MAC isoflurane (from 250 to 88 ml [center dot] 100 g sup -1 [center dot] min sup -1), but in exact proportion to the decrease in myocardial oxygen consumption, allowing unchanged myocardial oxygen extraction and similar endocardial-to-epicardial myocardial flow ratios. [198]
In neonatal rabbit hearts studied in vitro with 1.5% halothane, McAuliffe and Hickey found no change in steady-state levels of high- energy phosphates or intracellular pH, despite a 50% decrement in mechanical performance. [199] Significant uncoupling of oxidative phosphorylation cannot account for halothane's depressant effect on systolic function in the neonate.
Anesthesia and Systolic Function
The effects of the inhalational anesthetics in intact immature hearts have been evaluated in several studies. Although one study suggested that the apparent increase in hemodynamic depression in the young heart in human studies may be a result of differences in anesthetic uptake and distribution, [200] other studies indicate the increased hemodynamic impairment of the volatile anesthetics is predominantly a result of increased direct myocardial depression in immature hearts. [200–203] Cook et al., for example, showed that immature (aged 15 days) rats developed cardiovascular failure at significantly lower myocardial halothane concentrations than did older rats. [13]
Volatile anesthetics depress contractility primarily by limiting Ca2+ availability to the contractile apparatus. They alter transsarcolemma and sarcoplasmic reticulum Ca2+ flux with the net result that intracellular Ca2+ stores are depleted. [204–208] Halothane depresses contractility more than isoflurane. [51,209] It decreases peak intracellular Ca2+ concentration more than isoflurane, which may be a result of its greater depressant effect on sarcoplasmic reticulum function or on other mechanisms. [210–213] In isolated rabbit hearts, halothane is a more potent depressant of contractile function (measured by peak systolic and developed LV pressures and +dP/dtMAX) than isoflurane in newborns and in adults. [51] With isoflurane, there are no differential effects between age groups, whereas halothane causes greater depression in neonatal hearts. Other studies in rabbits have also found more depression by halothane of tension development in isolated RV tissue in newborns than adults and more inhibition of RV ventricular papillary muscle contractile function by halothane and isoflurane. [42,214] In this case, inhibition was similar for both anesthetics. Other investigators have found an age differential in depression of contractility by isoflurane in isolated right atrial tissues of rat, but this may not be comparable with rabbit ventricle. [215] This differential age effect may occur at many sites in the excitation-contraction mechanism, but little information regarding this is available.
Several studies in the chronic neonatal lamb preparation have shown that halothane [196] and isoflurane [197,198] decrease cardiac output to a similar extent as oxygen consumption. Isoflurane at 1 MAC primarily decreased blood pressure by decreasing cardiac output rather than decreasing systemic vascular resistance. [198] With hypoxemic stress, the isoflurane anesthetized lamb (1 MAC) responds by increasing cardiac output and oxygen delivery. [197]
The results of several studies in children, infants, and neonates, are somewhat variable, probably a result of differences in technique and the ability to use completely load-independent measures of contractility. Potentially confounding factors such as changes in heart rate and afterload make pure measurements of contractility difficult. Nicodemus et al., in an early study with halothane, showed an increased incidence of hypotension in neonates compared with infants and less in infants than older children. [12]
Given the increased myocardial depression of other Ca2+ blockers, e.g., verapamil, diltiazem, and nifedipine in newborns, [81–83] it might be expected that volatile anesthetics also depress neonatal more than adult sarcolemma Ca2+ channels. The differential effects between age groups with halothane, but not isoflurane, may be because isoflurane is a relatively weaker negative inotrope or may be a result of different actions of the agents that have not all yet been elucidated. [212] Because halothane has specific effects on the sarcoplasmic reticulum not found with isoflurane, this additional level of myocardial depression might be expected to be absent or less prominent in immature myocardium with its poorly developed SR. Krane and Su found that the newborn rabbit is less sensitive to halothane depression of SR Ca2+ uptake than adults but more sensitive to halothane-induced Ca2+ efflux, which could lead to a reduction in intracellular stores and contribute to greater depression of contractility. [213,216] Studies of skinned rabbit myocardial fibers have yielded conflicting results as to whether newborns are more sensitive than adults to depression of Ca2+ sensitivity of contractile proteins by halothane and isoflurane. [213,217]
Murray et al., using echocardiographic derived indices of cardiac function in a group of neonates and infants, could show no difference in myocardial depression using equipotent concentrations of halothane or isoflurane. [200] A recent study using load-insensitive measures of ventricular function found that sevoflurane depresses contractility less than halothane in young children, although that study did not evaluate young infants. [218]
Anesthetics and Diastolic Function
Anesthetic effects on Ca2+ flux that impair systolic function also impair diastolic function. In rabbit hearts in vitro, this effect (as measured by left ventricular diastolic pressure, -dP/dtmax, and tau, the time constant of isovolumic relaxation) is greater with halothane than with isoflurane, and with halothane, it is greater in infants than in adults. [51] Depression by isoflurane is similar between the age groups. The greater effect of halothane in infants may reflect the immature myocardium's relatively limited capacity to remove Ca sup 2+ from the contractile proteins. Sequestration in SR is the principal mechanism responsible for relaxation in mature myocardium, but immature hearts in which sarcoplasmic reticulum is relatively undeveloped are likely to depend more on sarcolemmal Na sup +-Ca2+ exchange. [16,219,220] Developmental changes in contractile protein isoforms and affinity for Ca2+ as mediated by the actin regulatory proteins may also affect anesthetic actions on relaxation, but this has not been examined. The differences between agents likely reflect their relative potencies in altering Ca2+ flux.
Anesthesia and Baroreflexes
Arterial baroreflexes are present in healthy and critically ill newborn humans. [221,222] There are conflicting reports regarding postnatal maturation of baroreflex control of heart rate. One study performed in conscious piglets reported that sensitivity for regulation of heart rate (change in heart rate per change in blood pressure) decreased with age. [223] Other studies with anesthetized animals have demonstrated that sensitivity either increased or did not change with age. [224–226] A decreasing sensitivity with age as demonstrated in the only study with conscious animals is consistent with a decreasing role for heart rate in the control of blood pressure in older animals.
Halothane depresses baroreflex control of heart rate at several levels of the reflex: central nervous system, autonomic ganglia, and the heart. [227–229] In young piglets, halothane reduced baroreflex sensitivity up to 80–90%, and the effect is dose-dependent up to 1 MAC. [230] Halothane also decreased resting heart rate and shifted the reflex heart rate response to a narrower range of blood pressure at lower heart rates. [230] Halothane is more depressant in younger than older animals when phenylephrine is injected to increase blood pressure and reflexively decrease heart rate. [230,231] When nitroprusside is injected to lower blood pressure and reflexively increase heart rate, halothane is equally depressant in younger and older animals. [230] In preterm human infants undergoing ligation of a patent ductus arteriosus with halothane anesthesia, a heart rate baroresponse is not observed with the alterations in blood pressure. [232]
Isoflurane also depresses baroreflex control of heart rate at multiple sites. [233,234] In human neonates, isoflurane decreased baroreflex sensitivity by approximately 70%. [235] Fentanyl similarly decreased baroreflex control of heart rate in human neonates by approximately 40–50%. [236] Nitrous oxide is also more depressant of the baroreflex in infant rabbits than in adults. [237]
Summary
Immature hearts are more profoundly affected by many anesthetics than are adult hearts. Maturational changes in a variety of cellular and subcellular systems and influences of the autonomic nervous system may be responsible, but as yet, specific mechanisms remain to be elucidated. Studies of the interactions of anesthetics with the immature human heart are for the most part lacking, and extrapolation from other species is difficult. This is a fertile area for investigation, particularly for studies of the effects of the newer clinical anesthetics, which are lacking.
The authors thank Drs. Carl Lynch, III, and David Teitel for their gracious reviews of the manuscript.