Hemodynamic problems are frequent in infants undergoing neonatal intensive care. The evaluation and treatment of such problems must take into account the developmental physiology of the neonatal cardiovascular system, cardiac structure, microstructure, and function, which in neonates differ in many important and fundamental ways from those of mature humans. For example, the myocardium of the newborn has a greater concentration of noncontractile elements, such as mitochondria, and an irregular orientation of the myofibrils. The neonatal myocardium utilizes glucose and lactate rather than the preferential metabolism of fatty acids of the mature myocardium. Calcium-induced calcium release, which marks the function of the sarcoplasmic reticulum, is absent in immature myofibrils.
As a result of its structural and metabolic immaturity, the neonatal heart is functionally limited; basal contractility is already close to maximal levels, and therefore there is little “contractile reserve,” and the neonatal myocardium is largely incapable of responding to further demands on its function. One important implication of this is the afterload sensitivity of the circulation. An increase in afterload commonly leads to a reduction in cardiac output. Many inotropic/vasopressor agents directly increase afterload and may therefore result in a decrease, rather than an increase, in systemic perfusion. Responses to all cardiovascular medications are affected by the metabolic immaturity, functional immaturity, and pharmacokinetic differences. Some drugs that are positive inotropes in the mature myocardium have negative inotropic effects on the immature one. Milrinone and other phosphodiesterase-3 inhibitors act by blocking the third fraction of phosphodiesterase; however, phosphodiesterases 3 and 4 are unbalanced in immature myocardium, and therefore blocking phosphodiesterase-3 may have relatively unpredictable effects in a neonate. Extrapolation from studies in older subjects is of little or no value, only studies investigating specifically neonatal populations are relevant.
Similarly, immature responses are also seen in the neonatal vasculature. We have little information about the development of the adrenergic vascular receptors, which are responsible for the responses to catecholamines. Vasoconstrictors, like dopamine, have been shown to increase systemic vascular resistance in the preterm newborn ; however, the stage of maturation at which such responses appear and the gestational age at which vasodilator responses may appear in response to other agents are unknown.
Another important factor that limits extrapolation of data from older patients is the presence of shunts. Both ductal and intracardiac shunts are frequent; there is therefore no single value for “cardiac output” in the sick neonate; left ventricular output, right ventricular output, systemic perfusion, and pulmonary blood flow are all potentially different numbers. In addition, interventions that differentially affect systemic and pulmonary vascular resistance may significantly affect systemic perfusion. Total systemic perfusion is equal to systemic venous return, that is, the sum of superior and inferior vena caval flow; pulmonary blood flow is equal to the sum of all of the pulmonary venous return to the left atrium. So in the absence of an intracardiac shunt, we have what seems initially to be paradoxical; systemic perfusion is equal to right ventricular output, whereas pulmonary blood flow is equal to left ventricular output; when there are significant shunts across the foramen ovale these statements have to be modified. Finally, left ventricular output is equivalent to systemic perfusion only when the ductus arteriosus has closed.
Pulmonary vascular resistance (PVR) is very high before birth, and less than 15% of the combined ventricular output perfuses the lungs during most of gestation; in human fetuses this proportion may increase prior to delivery at term. Right ventricular output mostly crosses the ductus arteriosus (from right to left) and perfuses the lower body and the low-resistance placental circulation. Right ventricular afterload is therefore low in utero; afterload increases at birth with clamping of the cord and then falls again as the PVR decreases with respiration.
Physiologic investigations have shown that clamping of the umbilical cord prior to initiation of breathing in a neonatal lamb model, when the PVR is still high, causes a reduction in left ventricular preload, in addition to the increase in right ventricular afterload. Delaying clamping until after breathing has commenced may avoid these changes, thus avoiding cardiovascular compromise around the time of birth. Further work confirming clinical benefits of delayed clamping in the very preterm infant is required.
Hemodynamic Problems in the Neonate
Persistent Pulmonary Hypertension of the Newborn
When PVR is persistently increased, or when it increases after an initial fall, there may be clinical consequences, known as persistent pulmonary hypertension of the newborn (PPHN). The most common underlying pulmonary disorders causing PPHN are meconium aspiration, septicemia, pneumonia, and pulmonary hypoplasia. In addition it can be seen occasionally in neonates with clear chest X-rays, as so-called primary or idiopathic PPHN.
The clinical presentation is hypoxic respiratory failure with the clinical signs of one of the underlying pulmonary conditions detailed above. In some patients this may be accompanied, if the ductus arteriosus is open, by a gradient in the saturations from pre- to postductal sites, indicating bidirectional or right-to-left ductal shunting, usually in the most severely affected patients.
On echocardiography many infants will be shown to have intracardiac shunting across a patent foramen ovale or they may have hypoxia from intrapulmonary shunting, that is, ventilation–perfusion mismatch. Intra-atrial shunting will occur when right atrial pressure is above left atrial pressure; right atrial pressures increase when right ventricular failure occurs, usually as a result of high right ventricular afterload. Studies have shown that right ventricular function is an important predictor of good outcome in infants with this condition.
Initial interventions should be supportive, including oxygen, fluid administration, warmth, minimal handling, and assisted ventilation. Infants who are agitated may benefit from sedation.
Oxygen should be given to achieve normal saturations, but hyperoxia should be avoided. Oxygen is toxic when given in higher concentrations, may increase pulmonary vascular reactivity, and may even decrease the response of the pulmonary circulation to nitric oxide. It does not appear that increasing FiO 2 beyond that required to achieve normal saturations has any effect on decreasing PVR. As for sedation, it is not clear which sedative agent is preferable. A hemodynamic study of infants with pulmonary artery catheters undergoing surgery showed that fentanyl reduces pulmonary vascular responses to endotracheal suctioning, which suggests that fentanyl may reduce pulmonary vasoreactivity and have a benefit in some infants with PPHN.
Hyperventilation should be avoided as it risks increasing pulmonary damage and causes cerebral vasoconstriction. Persistent respiratory alkalosis seems to cause progressive systemic hypotension, at least in some animal models. It also may cause potential adverse long-term neurodevelopmental effects and hearing loss. Bicarbonate should be avoided, as its use has been associated with an increase in mortality and an increased need for extracorporeal membrane oxygenation (ECMO). Optimizing lung inflation and ventilation to achieve a “normal” pH is reasonable, but going beyond this to alkalinize the patient, even if this might lead to a short-term improvement in P o 2 , is not supported by any evidence.
The only specific evidence-based therapy for PPHN is inhaled nitric oxide. Nitric oxide can be commenced at between 2 and 20 ppm ; there is little evidence that increasing beyond the initial concentration improves any clinical responses. More than 50% of children with PPHN will have a definite increase in the oxygen saturation after starting nitric oxide. Nitric oxide decreases the number of infants who will deteriorate to the point of needing ECMO; the number needed to treat to prevent one case of ECMO, among term newborns and late preterms with hypoxic respiratory failure who have reached an oxygenation index of 25, is 5.
Cardiovascular support including the use of inotropes may be required for infants with PPHN, but there is little evidence to support a choice of one agent over another. The most appropriate agent would have no pulmonary vasoconstrictor effects, or be a pulmonary vasodilator, one that would increase contractility and cardiac output without increasing vascular resistance. No agent is known to have all these effects.
In animal models dopamine increases both systemic vascular resistance (SVR) and PVR equally, unless enormous doses are used, suggesting that it may not be the best choice. In some animal models epinephrine has a greater systemic than pulmonary pressor effect. Norepinephrine also may be reasonable choice; a small observational study in full-term infants showed a good response to norepinephrine. Other agents such as milrinone and levosimendan have been suggested, and some animal models do show possible pulmonary vasodilatation with milrinone, but there is limited clinical research data to support their use. Evidence for the use of milrinone in PPHN is limited to case series demonstrating an improvement in oxygenation when used in infants failing to respond to inhaled nitric oxide. Milrinone is thus a potential candidate in this setting that deserves further investigation. If investigation of the hemodynamic responses in the sick newborn with PPHN confirms the potential inotropic, lusitropic, and pulmonary vasodilator responses, this should then be followed by investigation of the clinical outcomes.
Comparative studies of various agents among those who require hemodynamic support are needed. It will be important to determine which agents increase systemic pressures more than pulmonary pressures, which increase systemic perfusion, and most importantly if any choice of agent affects clinical outcomes.
The hemodynamic features of septic shock in the newborn have not been well described. Adults with gram-negative septic shock often present with so-called warm shock; this is a combination of excessive vasodilatation with incomplete cardiac response; cardiac output may be increased or within the normal range, and the patient often presents with hypotension; many of the changes are due to the endotoxins (in particular lipopolysaccharides) produced by the responsible organisms. Newborn infants with their different cardiovascular physiology and different bacteriology may present with a more variable profile. Newborn animals (such as piglets) with group B streptococcus more commonly have cold shock, with marked reductions in cardiac function (as a result of exotoxins produced by the organisms), and blood pressure is maintained initially with profound vasoconstriction, hypotension being a preterminal event. Some infants with Escherichia coli or other gram-negative sepsis seem to present with typical warm shock, but there are very few descriptions of the hemodynamics of sepsis in the literature. One study described hemodynamic features in several septic infants, but only a minority had signs of circulatory compromise or shock, as evidenced by the fact that several received neither fluid boluses nor inotropes. The organisms involved were variable; the infants tended to have low SVR and relatively high left- and right ventricular outputs. A more recent study of infants with septic shock suggested that the infants had mostly warm shock, but this study had a number of limitations.
If the above considerations are appropriate, we can divide the clinical presentations into warm and cold shock: Infants with sepsis and cold shock are vasoconstricted with prolonged capillary filling, adequate blood pressure, and oliguria. They are often lethargic and may have biochemical signs of poor oxygen delivery. Infants with warm shock on the other hand have bounding pulses, hypotension, and normal capillary filling, but may also be oliguric with lactic acidosis.
Evaluation of the circulatory status with echocardiography may well be important in such patients and may aid in targeted management strategies. Echocardiographic evaluation should include an analysis of cardiac filling, contractility, and systemic blood flow. Although such an evaluation may aid in providing more rational treatment, there is no clear evidence that it improves outcomes, and it is uncertain how echocardiographic findings should guide the selection of specific interventions.
A reasonable therapeutic approach is to use physiology-based medicine. This implies examining the abnormalities found on clinical evaluation combined with echocardiography. Therefore an infant with echocardiographic signs of reduced cardiac filling should receive a fluid bolus; an infant with reduced perfusion but adequate blood pressure may benefit from dobutamine or very low-dose epinephrine (which increases systemic perfusion with little effect on blood pressure). Infants with shock and hypotension may receive a moderate dose of epinephrine, which appears to increase both blood pressure and systemic perfusion. Norepinephrine has been little studied in the newborn, but one published study and our own experience suggest that it may have a very favorable hemodynamic profile.
The pharmacokinetics of the drugs is extremely variable; in addition, the concentration, affinity, and activity of the adrenoceptors are extremely variable. There is no consistent relationship between plasma catecholamine concentration and target-organ effect. Thus, in general, for individual catecholamine infusions, the hemodynamic response to therapy is not related to plasma concentrations in a simple linear fashion. The response to a particular plasma catecholamine concentration varies with the functional state and density of adrenergic receptors and the capacity of the target organs to respond. This means that dose responses are unpredictable and doses should be individualized.
Norepinephrine is now the first-line inotropic agent advocated in adult sepsis. However, a systematic review of data from adults with septic shock showed no difference in survival or other important outcomes from trials comparing inotropic agents, despite differences in short-term hemodynamic responses. There are no such trials of norepinephrine in the newborn.
Even in the absence of signs of inadequate preload, septic patients are often considered to have “functional hypovolemia.” They therefore receive fluid boluses, often multiple. However, a trial in older infants showed an increase in mortality in the group of children with early septic shock who were randomized to a fluid bolus, compared to controls who were not, calling into question this common practice. If we decide to give a fluid bolus, what fluid should we use? Acute responses to crystalloids and to colloids are different; the increase in systemic perfusion with colloids appears to be greater and more prolonged compared to saline. There is, however, no evidence that clinical outcomes are different, and several trials are currently under way to try to answer this question. The choice of fluid in the newborn is uncertain.
Further studies are clearly needed. Interventions for septic shock will probably need to be individualized according to the hemodynamic profile of the patient. Mortality from septic shock in the newborn is very high, so research in this area is clearly warranted. These studies should determine whether gram-negative and gram-positive shock have similar or differing profiles in the newborn. The role of bedside targeted neonatal echocardiography in the setting of neonatal sepsis needs to be investigated. The place of fluid-bolus therapy in the newborn needs to be evaluated, and the hemodynamic effects and clinical responses to various agents need to be evaluated. Finally the role and place of steroids, which are often given in treatment of septic shock, should be determined.
Infants with hypoxic–ischemic encephalopathy (HIE) may have a number of serious cardiovascular challenges, with myocardial insufficiency, leading to cardiogenic shock, as well as bradycardia, hypotension, and pulmonary hypertension. Infants with HIE often receive therapy with hypothermia, which leads to a further reduction in heart rate and blood pressure, and may be associated with a worsening of pulmonary hypertension. Infants undergoing hypothermia therefore are more likely to receive inotrope/vasopressor therapy, but it is not clear what the effects of hypothermia are on the pharmacokinetics and pharmacodynamics of the commonly used agents. Thresholds for (and goals of) intervention during hypothermia treatment are also not certain. Cardiovascular instability potentially leading to impairment of brain blood flow may contribute to adverse outcome, including mortality and adverse neurodevelopmental outcome. Therefore treatment aimed at preventing hypotension, poor myocardial contractility, and reduced cardiac output may have long-term benefits.
Cardiogenic shock is encountered most commonly after perinatal asphyxia. Other causes include following cardiac surgery or infants with hypoplastic left heart syndrome who may have profound shock, usually following closure of the ductus arteriosus. Aberrant coronary artery origins, although rare, should be considered in the absence of other clear etiology. Other causes such as cardiomyopathy and myocarditis are also possible but uncommon.
Usually such infants have poor perfusion and are often tachycardic (not always in the asphyxiated infant); increasing serum lactate, often leading to a frank acidosis, and oligoanuria may occur. This is one situation in which echocardiography is essential. As well as an analysis of cardiac function, the cardiac structure, including a verification of normal coronary artery distribution, should be examined.
As mentioned, even the healthy neonatal myocardium is intolerant of increases in afterload. When the primary problem is cardiac dysfunction it is essential to avoid increasing afterload. Agents that support cardiac function and decrease afterload, such as dobutamine and low-dose epinephrine, are reasonable first choices. Newer agents such as levosimendan, and perhaps milrinone, at least in the full-term infant, warrant further investigation. Excessive fluid administration should be avoided; even single fluid boluses should be carefully considered and given only if there is a good reason to suppose that there is hypovolemia.
Hypotension in the Extremely Low Gestational Age Newborn
In the first few days of life, preterm infants of less than 28 weeks’ gestation are often treated with fluid boluses and inotropes after a diagnosis of hypotension. One large prospective cohort study showed that, among infants born at 23 weeks’ gestation, 93% received a fluid bolus and over half were treated with an inotrope (usually dopamine). Even at 27 weeks 73% received a fluid bolus and 25% were treated with dopamine. There was huge variation between the 14 hospital centers, but the variation was not related to differences in patient characteristics, rather to variations in practice patterns. The most consistent finding of this study is that in the majority of circumstances intervention commenced on the first day of life (90%, 89%, 91%, and 89% of infants born at 23 to 24, 25, 26, and 27 weeks of gestation). This is a time period in which notable cardiorespiratory changes occur. Long-term outcome data from the same cohort show no evidence of benefit from more aggressive treatment of hypotension.
Many of these extremely immature babies are being treated with fluid boluses and inotropic agents despite there being no clinical or other objective evidence that they are underperfused. A number of studies have shown a poor correlation between indicators of systemic blood flow or oxygen delivery and mean arterial blood pressure in the preterm infant. Most hypotensive preterm infants have flows in the superior vena cava and/or right ventricular output that are within the normal range. The converse of this is that infants with low systemic flow may well have normal blood pressure, and thus patients who may benefit most from appropriate intervention are often underrecognized.
The combination of low numeric blood pressure with adequate systemic flow means that the SVR is low. On the first day of life this low SVR is unlikely to be due to an open ductus arteriosus, PVR is high, and shunting across the ductus is relatively limited for the most part. A low SVR with adequate oxygen delivery appears therefore to be part of the normal postnatal adaptation of the (very abnormal) extremely low gestational age newborn. Low numeric blood pressure without signs of poor perfusion may well not require any treatment; a small retrospective cohort study showed that good results can be seen with a permissive approach, avoiding active intervention for well-perfused infants who have low blood pressure. Treating normal transition may not be the best course of action.
Intervention algorithms and surveys of practice are characterized by a very similar approach, volume administration followed by dopamine. However, we need evidence whether treating low blood pressure as such in the preterm infant is beneficial and safe. Prospective randomized trials of intervention based on currently commonly used thresholds for intervention are warranted; some of these are under way as of this writing. The European Commission has funded two of these studies in an attempt to determine the efficacy of dopamine and dobutamine in the extremely preterm infant. The HIP trial is enrolling as of this writing to determine whether a standard approach to management of hypotension with volume and dopamine, versus a more observational approach, will result in an improved outcome. The Neocirc group will determine if dobutamine therapy results in an improvement in outcome in preterm infants with low flow. The TOHOP study is currently enrolling patients with low blood pressure and is attempting to evaluate the role of near-infrared spectroscopy as an adjunct in the management of low blood pressure. It is hoped that these studies will provide the clinician with evidence to direct therapy in preterm infants in the first few days of life.