Diagnosis and Treatment of Low Systemic Blood Flow in Preterm Infants

Abstract

After completing this article, readers should be able to: Low systemic blood flow (SBF) commonly occurs in preterm infants who have respiratory distress in the first postnatal day. Such low flows are strongly associated with neonatal morbidity, including peri/intraventricular hemorrhage (P/IVH), developmental impairments, and mortality. Diagnosing which infants have low SBF is problematic; use of blood pressure (BP) and signs of poor tissue perfusion, including capillary refill times (CRT), poor renal function, and acidosis, results in substantial delays in the initiation of cardiovascular support and a proportion of affected infants missing treatment. The traditional approach to cardiovascular support of the preterm infant has been to identify infants who have hypotension and to use agents, particularly dopamine, titrated to improve BP. However, this strategy may not be the best for improving SBF, and it is yet to be demonstrated that this approach improves neonatal outcomes.This review focuses on identifying which infants have low SBF or organ blood flow, discusses mechanisms for low flows, and examines potential therapies for treating or preventing low flow.Cardiovascular compromise occurs in a number of clinical scenarios in preterm infants, but it is most common on the first postnatal day. Not all infants who have low SBF will be suspected based on clinical signs, including criteria for hypotension. The first step in the approach to infant care is to recognize which infants are at risk of low flows: Three potential strategies for identifying infants in need of cardiovascular support include use of: The strategy employed depends on whether the neonatologist has developed skills in functional echocardiography of the newborn. To date, no specific approach to diagnosing cardiovascular compromise has been proven to improve clinical outcomes of infants. However, clearly there are substantial limitations to the traditional approach of identifying infants who have hypotension and using interventions designed to improve BP and not necessarily organ blood flow.Most trials of cardiovascular support in preterm infants have enrolled infants who have hypotension or clinical signs of poor tissue perfusion. This approach may result in substantial delays in treatment of many infants and fails to treat some infants who have low SBF. Hypotension is commonly defined as a BP less than the 10th percentile for birthweight or gestation, for which a common approximation used in the first postnatal days is a mean BP (mm Hg) below the gestation in weeks. Although the ranges of BP found in preterm infants during the first postnatal day are reasonably well defined, few data define physiologically normal BP in terms of organ blood flow or absence of organ dysfunction. In fact, there is a poor correlation between BP and blood flow measured as right ventricular output (RVO) (1) and brain and upper body blood flow (superior vena cava [SVC] flow) (2) during the first day (Fig. 1). This is hardly surprising because BP is the product of flow and vascular resistance (VR).In a cohort of 128 infants, Osborn and associates (2) found that 44 (34%) had low SVC flows (<41 mL/kg per minute) in the first day, with all infants who developed low SVC flow having done so by 10 hours. Although correlated with SVC flow, both BP and CRT were imperfect predictors of low flow (Table 1). A CRT of at least 3 seconds had 55% sensitivity (Sn) and 80% specificity (Sp) for low SVC flow, a mean BP lower than 30 mm Hg had 59% Sn and 77% Sp, and a systolic BP lower than 40 mm Hg had 76% Sn and 68% Sp for detecting low SVC flow. Combining a mean BP lower than 30 mm Hg and a CRT of at least 3 seconds increased the Sn to 78%.In a previous cohort of 126 infants undergoing blinded echocardiographic monitoring, only 31% who had low flows at 12 hours after birth had received inotropic support on the basis of hypotension; 82% were receiving inotropes by 24 hours. (3) This represents a considerable delay in treatment from the time of onset of low flow, considering that low flow usually occurs within the first 12 hours after birth. In addition, infants who have hypotension after the first day with ongoing inotrope requirements most frequently have cardiac outputs in the normal or high range. (4)In most adults and children, RVO and left ventricular output (LVO) are equivalent and represent actual SBF. However, measuring SBF in the preterm infant in the first day after birth is problematic because of the presence of significant shunts across the foramen ovale and ductus arteriosus (DA) in the adapting heart that result in the potential to overestimate SBF by up to 100%. (5)(6) (See other articles in this issue.) Two potential alternatives to ventricular outputs as a measure of SBF are cardiac inputs from the brain and upper body (SVC flow) (7) and organ blood flows using Doppler ultrasonography, near-infrared spectroscopy (NIRS), or xenon clearance techniques. Of these, measuring SVC flow via ultrasonography has the greatest clinical potential. SVC flow can be measured serially using this technique at the bedside of even very sick, extremely preterm infants by a neonatologist trained in echocardiography, which is becoming increasingly common. (8) SVC flow is not affected by DA or foramen ovale shunts and is a strong predictor of neonatal mortality, late P/IVH, (9) and neurodevelopmental impairments. (10) (See the article on low systemic blood flow and brain injury in this issue.) Almost all infants who have a late P/IVH have low SVC flow in the first postnatal day. (9)An alternative to measuring cardiac inputs (SVC flow) is to measure organ blood flow, most importantly, cerebral blood flow. Doppler ultrasonography has been used to measure organ blood flows in preterm infants, but because the diameters of these vessels are small, they are difficult to measure accurately. As a result, surrogates of flow have been used, including the pulsatility index and mean or maximal flow velocities, with changes in these taken to represent changes in absolute flow. Evans and associates (11) performed serial Doppler measurements in 126 infants born at fewer than 30 weeks’ gestation. They examined middle cerebral artery (MCA) size, velocities, pulsatility index, and estimated flow as well as SVC flow and BP. In multivariate analysis, after controlling for gestation, there was a highly significant association between lowest SVC flow and subsequent P/IVH, but no association between P/IVH and lowest MCA mean velocity, estimated diameter, pulsatility index, or mean BP. This finding suggests that BP and peripheral arterial Doppler ultrasonography are unlikely to be adequate measures of organ perfusion and that SVC flow in the first day and subsequent ventricular outputs in infants who do not have significant shunts are required to determine which infants have low SBF and are at risk of organ injury. Alternatively, cerebral blood flow can be measured using NIRS or xenon clearance techniques, although these techniques have limitations that restrict their application in clinical practice. To date, there are few studies examining cardiac outputs and these measures of cerebral blood flow and no studies in infants who have low SBF. Importantly, however, Meek and colleagues, (12) using NIRS to measure cerebral blood flow in the first day, confirmed that infants who develop a late P/IVH have a period of low cerebral blood flow preceding the hemorrhage.In two prospective cohorts totaling 254 infants, low SVC flow (<41 mL/kg per minute) was documented in more than one third of infants born at fewer than 30 weeks’ gestation. (9) Preterm infants who developed low SVC flow in the first day were of lower gestation (Fig. 2) and had more severe respiratory distress. Nearly all were mechanically ventilated. In the second cohort, of those infants who developed low flow, one third had done so by 3 hours of age; nearly all infants in both cohorts had developed low flow by 5 to 10 hours. Low flows were uncommon after 24 hours. The first hours after birth are a period of rapidly falling flows in many infants. As will be discussed later, despite standard treatments for low SVC flow, including early administration of dopamine and dobutamine, nearly 40% of affected infants fail to maintain “normal” SVC flow and are at very high risk of P/IVH and mortality.Instead of treating infants after they develop low flow, an alternative is to identify infants at risk of low flow and initiate treatment to prevent low flow, a strategy we currently are using in a trial of milrinone (an “inodilator”). By selecting all infants born at fewer than 28 weeks’ gestation and infants born at 28 to 29 weeks’ gestation whose SVC flows are less than 60 mL/kg per minute at 3 hours, cardiovascular support can be targeted to infants at risk of developing low SVC flow with 96% Sn and 69% Sp. This is considerably superior to using BP or CRT as indicators of the need for treatment. Where echocardiographic facilities are not available, selection of all infants born at fewer than 28 weeks’ gestation and infants born at 28 to 29 weeks’ gestation who are ventilated and have mean airway pressures (MAP) greater than 8 cm H2O and inspired oxygen levels greater than 30% at 3 hours of age has similar diagnostic accuracy.In multivariate analysis of the previously noted two cohorts of 254 infants that was corrected for multiple perinatal variables, risk factors for low SVC flow were remarkably consistent and included lower gestation, higher average MAP in the first 12 hours, and DA diameter greater than 1.6 mm at 5 hours (first cohort only). (9) In both cohorts, although SVC flow was correlated poorly with BP, low SVC flow was associated strongly with a high calculated upper body systemic VR. This raised the question of whether infants who have low flow have poor myocardial contractility in response to high left ventricular (LV) afterload. In the second cohort, we measured LV stress and mean velocity of circumferential fractional shortening (mVcfs) of the left ventricle. The slope of the relationship between mVcfs and LV stress describes left ventricular contractility, that is, its ability to contract against a given afterload. Among infants who developed low SVC flow, there was no significant difference in mVcfs or LV stress, but there was a significant increase in the slope of the relationship at 3 hours of age (Fig. 3). Immature infants who have low SVC flow in the first day have poor myocardial contractility, which may not be improved by inotropes.Low cardiac outputs and hypotension are common in infants who have sepsis. Cardiovascular compromise can occur with some infants who develop low-output states associated with high systemic VR (“nonhyperdynamic” sepsis) and others, particularly those who have gram-negative infection, who develop high-output states characterized by low systemic VR (“vasodilatory shock”). In addition, a substantial capillary leak can occur with subsequent hypovolemia. Unfortunately, most of this information is derived from infants out of the newborn period; little confirmatory data are available in preterm infants. However, our clinical experience is consistent with these observations, and we have observed that pulmonary hypertension can complicate sepsis, especially early-onset group B streptococcal infection.Low-cardiac output states occur in newborns who have perinatal asphyxia. (13) Such infants have reduced myocardial contractility (14) and variable degrees of pulmonary hypertension. (13)Newborns who have high oxygen and ventilator requirements with or without pulmonary hypertension frequently have low cardiac outputs. (13) Among preterm infants, the incidence of low ventricular outputs (particularly RVO) in the first days after birth increases with worsening respiratory distress. (5) The incidence of right-to-left ductal shunting indicating severe pulmonary hypertension was significantly increased in infants who had fatal respiratory distress syndrome. (5)The usual principle of caring for infants who have evidence of cardiovascular compromise is to identify the type of infant being managed and to determine likely underlying mechanisms. This is facilitated by rapid and timely echocardiographic assessment. Useful information includes presence, size, and direction of a ductal shunt; ventricular outputs; and SVC flow if the infant has a ductal or atrial shunt. This allows identification of the hemodynamic significance of the DA, determination of the presence and severity of pulmonary hypertension, and an estimate of SBF. Where echocardiography is not available, interventions are directed toward infants who have clinical evidence of cardiovascular compromise (hypotension or prolonged CRT). Potential therapies for infants who have low cardiac outputs include volume expansion; inotropes such as dopamine, dobutamine, and adrenaline; nitric oxide (NO); corticosteroids; and “inodilators” such as milrinone.Despite evidence from meta-analysis of randomized trials that prophylactic indomethacin prevents P/IVH, (15) there is insufficient proof to determine if the use of cyclooxygenase inhibitors to close a large DA in the first hours after birth results in improved or maintained SBF. Among infants who had symptomatic DAs after the first postnatal day, indomethacin but not ibuprofen reduced cerebral blood flow. (16) However, few data are available for the first day in infants who have large DAs. In a randomized, crossover trial of indomethacin in infants born at fewer than 30 weeks’ gestation who had large DAs in the first hours, indomethacin had minimal effect on DA constriction at 1 hour compared with placebo and little effect on SVC flow or RVO. (17) In uncontrolled measurements, significant DA constriction occurred 2 hours after indomethacin. At this time, SVC flow was relatively well maintained, when previous data suggest that it should be falling. Further studies of targeted indomethacin in preterm infants who have large DAs in the first hours are required to determine the effect of this agent on SBF.There is no evidence from randomized trials that routine use of volume expansion in the first day in preterm infants improves any neonatal outcome, including mortality, P/IVH, or subsequent neurodevelopment. (18)(19) However, most trials comparing volume expansion with no treatment did not enroll infants on the basis of cardiovascular compromise, making the effect of volume on SBF in preterm infants who had hypotension or low SBF uncertain. Where infants who had cardiovascular compromise were enrolled, two different types of volume expansion (18) or volume versus inotrope (19) were compared. In one small trial of hypotensive infants, albumin was not as effective as dopamine for correcting hypotension, but there was no significant difference in short-term outcomes, including mortality, P/IVH, and periventricular leukomalacia (PVL). (20) Two trials (one unpublished) compared the use of colloid (albumin 5%) with crystalloid (saline) in hypotensive preterm infants. So and colleagues (21) reported no significant difference for any neonatal outcome, including the incidence of failed treatment (persistent hypotension). In an abstract, Lynch and associates (22) reported that 10 mL/kg albumin administered to hypotensive preterm infants in the first 3 days resulted in a significantly greater increase in BP than saline and a reduced requirement for pressor support. Effects on blood flow were not reported.In observational studies, Osborn and associates (23) found that 10 mL/kg saline administered to infants who had low SVC flows produced a mean 43% increase in SVC flow and 22% increase in RVO, with 55% of infants increasing flows above 40 mL/kg per minute. Because all infants then received inotropes, the longer-term effects of volume are uncertain. The place of volume expansion in the preterm infant remains unclear. A high index of suspicion is required for hypovolemia in the clinical setting of peripartum blood loss, particularly when there is a tachycardia and pallor. If the blood loss is acute, an urgently obtained hematocrit measurement still may be normal. Outside this scenario, hypovolemia does not appear to be a substantial problem in most preterm infants, with studies demonstrating a poor correlation between hypotension and blood volumes in preterm infants. In view of the substantial short-term increase in SVC flows seen in many infants, judicious use of volume (saline 10 to 20 mL/kg) should be considered. Larger volumes should be avoided unless the infant is likely to have hypovolemia.Among the inotropes and vasopressors that have been used in preterm infants are dopamine, dobutamine, adrenaline, and isoprenaline. The mechanisms of action of these agents often are complex and affected by the developmental maturation of the cardiovascular and autonomic nervous systems in the preterm infant. Unfortunately, nearly all studies of inotropes in preterm infants have enrolled infants on the basis of hypotension, resulting in the evaluation of many infants who have normal systemic and organ blood flows. It had not been recognized that many extremely preterm infants suffer from low SBF in the first postnatal day, but there now is evidence that a substantial proportion have low SBF and are at risk of organ damage. (3)(5)(9)(10)(11)(12) Any analysis of the effect of treatment on blood flow must take into account the problem of shunts across the adapting heart, which make the use of ventricular outputs, particularly LVO, inaccurate for assessing SBF. Doppler measurement of SVC flow and direct measurements of organ blood flow are the current best measures of blood flow in preterm infants in the first day. A major limitation of any review is that, to date, there are no data on long-term neurodevelopmental outcomes from inotrope trials.Dopamine is a naturally occurring catecholamine precursor of noradrenaline and adrenaline that can stimulate alpha, beta, and dopamine receptors. Dopamine infusions produce substantial increases in the levels of adrenaline and noradrenaline, which mediate some of its cardiovascular effects. Low levels of noradrenaline and reduced sympathetic innervation in the immature myocardium may modify the effect of dopamine in the preterm infant. The effects of dopamine are dose-dependent. At infusion rates of 0.5 to 2 mcg/kg per minute, dopaminergic receptors are stimulated, resulting in renal and mesenteric vasodilatation and little change in BP. At 2 to 10 mcg/kg per minute, beta1 receptors are activated, and cardiac output and systolic BP increase. At doses greater than 10 mcg/kg per minute, alpha receptors are activated, causing vasoconstriction and increases in systolic and diastolic BPs. Dopamine has been used extensively in neonates for treatment of hypotension. Dopamine also has noncardiovascular effects, suppressing prolactin, thyrotropin, and growth hormone secretion.Several studies have examined the effects of low-dose dopamine compared with no treatment to improve renal function in preterm infants, including ventilated infants who had respiratory distress syndrome (RDS) (24) or infants receiving indomethacin. (25) In general, relatively little benefit in terms of renal function and no improvement in clinical outcomes have been demonstrated, so this therapy currently cannot be recommended. Randomized trials of hypotensive preterm infants in the first days after birth have compared dopamine with volume expansion, dobutamine, and adrenaline.Dobutamine is a synthetic catecholamine that has beta-adrenergic and cardiac alpha-adrenergic effects. Dobutamine increases myocardial contractility by direct stimulation of myocardial alpha-adrenergic receptors, and unlike dopamine, its pharmacologic action is not dependent on peripherally released noradrenaline. Dobutamine has been used in asphyxiated hypotensive preterm infants who had myocardial dysfunction (26) and compared with dopamine in several randomized trials in hypotensive preterm infants. (27) The dose range used in clinical trials has been up to 20 mcg/kg per minute. There are no published randomized, controlled trials (RCTs) of dobutamine compared with placebo or no treatment in newborns. Several observational studies have reported cardiovascular effects of dobutamine in newborns. Devictor and colleagues (26) reported that at 10 mcg/kg per minute, dobutamine produced significant increases in cardiac output, heart rate, and aortic blood flow velocity in six term infants who had severe perinatal asphyxia. Martinez and associates (28) reported a significant increase in cardiac output but no significant change in BP or heart rate at infusion rates of 5 and 7.5 mcg/kg per minute in sick newborns born between 27 to 42 weeks’ gestation. Stopfkuchen and coworkers (29) enrolled 17 ventilated preterm infants on the basis of an elevated ratio of pre-ejection period to left ventricular ejection time (PEP/LVET) that reflected abnormal LV function. They reported that a dobutamine infusion resulted in a significant increase in heart rate, decrease in PEP/LVET ratio, and increase in mean BP, suggesting an improvement in left ventricular contractility.Adrenaline is a naturally occurring sympathomimetic amine that has both alpha- and beta-adrenergic effects. It is used widely in the context of cardiorespiratory arrest and cardiovascular support across all age ranges. There are few pharmacokinetic data regarding adrenaline in preterm infants, but in sick children, infusions of 0.03 to 0.2 mcg/kg per minute of adrenaline have cleared rapidly, producing half-lives measured in minutes. Adrenaline infusions are used most frequently in neonates for the treatment of hypotension refractory to other inotropes. (30)(31) Trials of adrenaline versus dopamine infusion have yet to be published.Most trials comparing dopamine with dobutamine have enrolled infants on the basis of systemic hypotension according to population-defined criteria, including systolic BP less than 40 mm Hg, mean BP less than 30 mm Hg, or mean BP less than the 10th percentile for postnatal age. (27) Both inotropes were used as continuous infusions titrated against BP, with a dose range of 5 to 10 mcg/kg per minute. Meta-analysis of the trials found that dopamine was associated with a reduced risk of treatment failure (defined as failure to improve BP into a defined range), but there was no difference in any clinical outcome, including mortality, P/IVH, PVL, and necrotizing enterocolitis (Table 2). (27) Only one study (32) reported cardiac outputs, noting that dopamine produced a mean 14% reduction and dobutamine a mean 21% increase in LVO (significant difference). In contrast, dopamine produced a significantly greater increase in mean BP than dobutamine associated with a significantly greater increase in systemic VR. This study did not report the presence and size of a DA shunt, which can make the use of LVO an inaccurate measure of SBF. Nearly all infants in the study had LVOs in the normal range.Only one study has enrolled infants on the basis of low SBF. Osborn and colleagues (23) evaluated 42 infants born at fewer than 30 weeks’ gestation who had SVC flows less than 41 mL/kg per minute in the first hours after birth. Infants received 10 mL/kg normal saline and were allocated randomly to dopamine or dobutamine. Inotrope infusions were initiated at 10 mcg/kg per minute and increased to 20 mcg/kg per minute if the SVC flow failed to increase or was not maintained above 40 mL/kg per minute. Dopamine produced a mean 4- to 5-mm Hg increase in mean BP at both infusion rates. Dobutamine produced little change in BP or heart rate. There were trends toward greater improvements in SVC flow with dobutamine at both infusion rates, with a significantly greater increase in SVC flow at the highest dose of inotrope reached compared with dopamine (mean increase +35% versus −1%). Overall, 68% of infants maintained SVC flow above 40 mL/kg per minute on dobutamine compared with 45% on dopamine (not significant). There were no significant differences for any clinical outcome between infants randomized to dobutamine or dopamine. Although rates of late P/IVH (18% versus 45%, P=0.06) and late grade 3 to 4 P/IVH (5% versus 30%, P=0.04) were lower among infants randomized to dobutamine, overall rates of P/IVH and grade 3 to 4 P/IVH were not significantly different (Table 3). Of concern was that 40% of infants failed to maintain SVC flows in the normal range, despite the use of both inotropes. These infants were less likely to have received a complete course of antenatal steroids and were more likely to have RDS. There was no evidence that myocardial contractility was increased by either inotrope, as evidenced by the relationship between mVcfs and LV stress. However, dopamine at an infusion rate of 20 mcg/kg per minute produced significantly increased LV stress. The adverse effect of dopamine on RVO in an infant can be seen in the attached files ( Video #1 ) ( Video #2 ) ( Video #3 ). For the infants receiving a single inotrope at 24 hours after birth, dobutamine produced a significantly greater RVO than dopamine.In newborn animal studies, both dopamine and adrenaline increased cardiac output, BP, and systemic and pulmonary VR. Adrenaline produced greater increases in systemic than pulmonary arterial pressures and resistance; dopamine, especially at higher doses, increased pressures and resistance similarly in both circulations. Two studies in preterm infants comparing dopamine and adrenaline have been published in abstract form. Phillipos and colleagues (33)(34) randomized infants to dopamine (starting at 5 mcg/kg per minute and increased to 10, 15, and 20 mcg/kg per minute) or adrenaline (starting at 0.125 mcg/kg per minute and increased to 0.25, 0.375, and 0.5 mcg/kg per minute until the desired mean BP was achieved for 1 hour). For infants whose birthweights were more than 1,750 g, a significant increase in heart rate and BP at the highest dose reached and study endpoint were reported for both adrenaline and dopamine. The differences between the two inotropes were not reported as statistically significant. Neither inotrope produced a significant change in LVO or RVO. Dopamine produced a significant fall in LV stroke volume compared with baseline that was associated with an increased systemic VR. No data were available for clinical outcomes, success of treatment of hypotension, or incidence of low cardiac output.Isoprenaline (isoproterenol) is a synthetic sympathomimetic amine that is structurally related to adrenaline and acts almost exclusively on beta-adrenergic receptors. In animals, isoprenaline infusion increases cardiac output but causes significantly greater increases in heart rate to achieve similar changes in cardiac output compared with dopamine and dobutamine. Systemic mean BP is decreased. Renal artery blood flow is increased by dopamine, unchanged by dobutamine, and decreased by isoprenaline. There are few published data on the use of isoprenaline in preterm infants, although it has been used clinically in infants who had congenital heart block, status asthmaticus, meningococcal septicemia, and persistent pulmonary hypertension of the newborn.NO is a selective pulmonary vasodilator administered as an inhalational gas in the ventilator circuit. This raises the possibility of selectively reducing pulmonary VR to increase cardiac outputs. Supporting this possibility are the findings that pulmonary hypertension is associated with severe and fatal RDS in preterm infants (5) and low ventricular outputs are common in term infants who have pulmonary hypertension. (13)(35) One study awaiting publication has examined the effect of NO in preterm infants who have RDS. (35) The investigators randomized 207 infants to NO or placebo in the first hours after birth and reported a significant decrease in the incidence of bronchopulmonary dysplasia and death (odds ratio [OR], 0.58; 95% CI, 0.33 to 1.00) and severe P/IVH and PVL (OR, 0.42; 95% CI, 0.19 to 0.92) with NO. These findings raise the possibility that reducing pulmonary VR with NO in the first postnatal day has the potential to prevent cerebral injury in preterm infants.Rationales for using corticosteroids to treat hypotension in infants include a probable reduction in adrenergic receptors and sympathetic innervation of the immature myocardium, which may reduce responsiveness of preterm infants to catecholamines, with glucocorticoids increasing numbers of beta receptor in in vitro experiments, as well as a relative adrenal insufficiency that has been reported in infants who have hypotension that is resistant to inotropes. Two RCTs of postnatal corticosteroids have documented increases in systemic BP and reduced need for inotropic support in hypotensive preterm infants. (36)(37) Among infants who had hypotension that was refractory to volume and inotropes, hydrocortisone produced a significant increase in BP within 2 hours and a reduction in inotropic support by 12 hours. (38) However, there was no measure of cardiac output or organ blood flow and no control group. In a systematic review of randomized trials of early postnatal corticosteroids for preventing chronic lung disease, the use of steroids in ventilated preterm infants resulted in a reduction in the incidence of PDA and increase in the incidence of hypertension. (39) There was no significant difference in mortality or rates of P/IVH, PVL, or necrotizing enterocolitis. Incidences of hypotension, low cardiac output or organ blood flow, and use of inotropes were not reported. One RCT compared the use of dopamine 5 mcg/kg per minute, titrated up to 20 mcg/kg per minute, with hydrocortisone 2.5 mg/kg in hypotensive preterm infants in the first 7 postnatal days. (40) There was no significant difference in treatment failure rate (absence of hypotension) among infants receiving dopamine compared with hydrocortisone (0% versus 19%, P=0.1), mortality (5% versus 10%), necrotizing enterocolitis (5% versus 19%), or P/IVH grades 2 to 4 (16% versus 24%). Measures of cardiac output and organ blood flow were not reported.Milrinone is a bipyridine compound that is a selective phosphodiesterase inhibitor. It is positively inotropic and reduces pulmonary and systemic VR. We have conducted a pilot study of milrinone infusion for prevention of low SVC flow in 29 infants who were at high risk of low SVC flow in the first 24 postnatal hours. Infants were enrolled if they were born at fewer than 27 weeks’ gestation. Among 10 infants who received an optimized dosing regimen from 3 hours (milrinone 0.75 mcg/kg per minute for 3 h, then 0.25 mcg/kg per minute for 15 h), none developed low flows compared with 61% in a historical cohort. We now are comparing milrinone with placebo for prevention of low flows in preterm infants at high risk of low SVC flow based on gestation and respiratory support criteria.Among infants at high risk of low SBF, there is a strong argument for measuring both blood flow using ultrasonography and BP. This allows the determination of VR, which has the potential to direct therapy. Using ultrasonography, SVC flow is the best current measure of SBF in the first day, with RVO measured as a cross check. When it is documented that there is no DA or foramen ovale shunt, ventricular outputs may be used. We propose a model of cardiovascular support in the preterm infant that recognizes the importance of low SBF, poor myocardial contractility, and high VR in the first day (Table 4 and Fig. 4). Hypotension, low VR, and normal or high SBF dominate beyond this period. Low-output states have been documented in infants who have severe RDS, pulmonary hypertension, asphyxia, and sepsis. Poor myocardial contractility occurs predominately in ventilated extremely preterm infants on the first postnatal day and infants who have perinatal asphyxia. Specific consideration must be given to the likelihood of hypovolemia, which is uncommon except in the context of peripartum blood loss and among infants who have sepsis and capillary leak (probably uncommon) or other sources of fluid loss, including gastrointestinal obstruction and some infants who have necrotizing enterocolitis or incarcerated gut.These infants have a high incidence of low-flow states with normal or low BP. Dobutamine, which may increase myocardial contractility, reduces systemic VR, and is more likely to increase SBF, is the most appropriate first-line therapy. It is logical to initiate treatment as early as possible on the basis of echocardiographically detected low SVC flow or RVO in the first hours or clinical suspicion based on low gestation, need for mechanical ventilation, and hypotension or prolonged CRT. Judicious volume expansion (normal saline 10 to 20 mL/kg) is reasonable, especially because dobutamine is a vasodilator. Second-line therapy at present should be dopamine starting at 5 mcg/kg per minute and titrated to attain an adequate SBF or minimally acceptable BP, depending on the availability of echocardiography. Doses in excess of 15 mcg/kg minute should be avoided because they raise systemic VR and have deleterious effects on SBF. For infants who have low SBF that fails to respond to these measures, adrenaline is likely to raise BP and may increase SBF. For infants who have normal SBF but hypotension, dexamethasone 0.25 mg/kg (36) or hydrocortisone 2 to 6 mg/kg (38) is likely to improve BP over the next 6 hours. Infants who have low SBF and fail to respond to inotropes are at high risk of mortality and morbidity.These infants may have low SBF with or without hypotension associated with poor myocardial contractility. Dobutamine is likely to be the most effective first-line therapy, with consideration given to adrenaline as second-line treatment among those who fail to respond.Infants who have hypotension beyond the first day likely have normal SBF and low systemic VR. Appropriate first-line treatment is dopamine starting at 5 mcg/kg per minute titrated to a minimally acceptable BP. It is uncommon for infants to fail to respond to dopamine unless there is an element of adrenal suppression. Corticosteroids, after a short adrenal stimulation test, may be reasonable second-line therapy for the infant who has normal SBF and persistent hypotension or who remains dependent on inotropes. Suggested approaches to infants who have sepsis (“hyperdynamic” and “nonhyperdynamic”) and pulmonary hypertension are listed in Table 4.Evidence suggests that extremely preterm infants suffer low SBF in the first day after birth that is associated with high VR and poor myocardial contractility. Use of BP and population criteria for hypotension to determine which infants need support may delay treatment for low SBF substantially and miss some infants who have low SBF. It has yet to be proven that targeting BP improves the outcomes of preterm infants. Ultrasonography provides the opportunity for rapid and timely measurement of SBF in preterm infants at risk of low flows. Vascular resistance can be calculated based on BP and SBF, and cardiovascular support can be directed more appropriately. To date, only one study has examined preterm infants with low SBF in the first day after birth. (23) In this study, dobutamine, which decreases VR, increased blood flow better than dopamine, which increases VR, especially at higher doses. Research now should target the treatment or prevention of low-flow states in the extremely preterm infants in the first postnatal hours. Ultrasonography provides the neonatologist with the ability to obtain rapid and timely serial measurements of ventricular outputs, SVC flow, DA patency and size, and an estimation of pulmonary hypertension. Care must be taken in interpreting ventricular outputs in the presence of a significant shunt (common in the first day). The use of SVC flow offers a satisfactory estimation of SBF that is easy to measure and has the ability to predict adverse outcomes. The next challenge in newborn care is effective direct treatment for infants who have low SBF and increasing these flows.