With the advent of routine early ultrasonography, the gestational age is fairly well established on admission to the Obstetrics unit for the vast majority of women presenting with premature labor, premature rupture of membranes or other problems diagnosed in the second trimester of pregnancy. Such patients need to be immediately placed in the charge of a specialist in high risk Obstetrics (Obstetrical Perinatologist or Specialist in Maternal Fetal Medicine) to coordinate the evaluation and management plans and to ensure appropriate communication and counseling. Based on the investigation for the causes of the problem at hand, the evaluation of the degree of cervical dilatation, the condition of the membranes, the presence or absence of chorioamnionitis and, if possible, the most recent evaluation of fetal well-being by bedside or formal ultrasonography, the Perinatologist can decide on the best course to be taken, such as an estimate of the likelihood of controlling labor with tocolysis to allow adequate time for antenatal corticosteroid therapy (7,29), often with consultation with their colleagues in neonatology.

The prospective parents should receive accurate information regarding all facets of the proposed management, including the possible need for cesarean delivery, and information regarding the subsequent management of the newborn infant, including the potential risks related to both the degree of prematurity and the therapeutic interventions that may be necessary to keep the infant alive.

Ideally, this information should be provided conjointly by both the obstetric perinatologist and the neonatologist and should be based not only on general statistical information, but also on the specific institutional experience with outcomes of newborn infants of similar gestational age. In our center, the attending neonatologist provides a written consultation on all patients admitted to the obstetric high-risk unit. We meet with the family, offer an extensive review of our experience with similar cases, and answer their questions regarding risks and outcomes. The father and mother are invited to visit the NICU and to familiarize themselves with the environment and the personnel. If the mother is not able to visit for medical reasons, we show her a book with pictures explaining each step of the baby’s treatments, from the delivery room to the time of discharge.

The lowest gestational age at which resuscitation should be initiated has long been the subject of debate. Guidelines are available from both the American and the Canadian Fetus and Newborn Committees (30,31). Based on our experience, we offer an optimistic opinion in terms of survival and potential morbidity for pregnancies of 25 weeks’ gestation and over. Between 24 and 25 weeks, although we underline that the chances of survival are quite good, we also emphasize the increased risk of potential complications, such as intraventricular hemorrhage (IVH), periventricular leukomalacia (PVL), retinopathy of prematurity (ROP), chronic lung disease (CLD), neurosensory impairments, and later school and behavioral difficulties. For pregnancies between 23 and 24 weeks of gestation, we describe the higher incidence of complications previously mentioned, and the lower survival rate; however we also mention the possibility of intact survival or survival with minimal handicaps. Finally, for pregnancies below 23 weeks of gestation, we do not recommend intervention. For parents who request full intervention, we strongly advise that resuscitation will be undertaken only if the newborn has at least the degree of maturity predicted by dates and/or ultrasonography, and if, in the judgment of the neonatologist present in the delivery room, the newborn has reasonable chances of responding to resuscitation. We always make it clear to the parents that initiation of resuscitation and subsequent treatments in the NICU do not preclude discontinuation of therapy if a major complication such as severe IVH is detected in the hours or days following birth. The presence of a staff neonatologist in the delivery room is an integral part of our protocol for the management of ELBW infants.

One of the most difficult questions that parents ask, and which our obstetric colleagues continuously debate, is the safest route of delivery in the presence of either a breech presentation or evidence of fetal distress (32). In our institution, based on our own results, we advise cesarean delivery in such situations as of 25 weeks’ gestation. Between 24 and 25 weeks, we feel less inclined to recommend cesarean delivery, particularly in view of the fact that many may require a classical incision. The decision to proceed with such an intervention is taken with a clear understanding by the parents of all the medical implications for both the mother and the infant. Finally, at less than 24 weeks of gestation, cesarean delivery is performed strictly for maternal indications. Our incidence of cesarean section by gestational age is indicated in Table 25-10. It is obvious that a large number of cesarean deliveries, and particularly those between 22 and 24 weeks of gestation, are performed strictly for maternal indications, i.e., severe abruption, preeclampsia, etc.

Another difficult management situation relates to ruptured membranes between 18 and 22 weeks of gestation, resulting in severe oligohydramnios, with the inherent risk of lung underdevelopment (33,34). Serial ultrasound studies can evaluate the degree of reaccumulation of amniotic fluid and allow for a better-educated decision regarding whether continuation of pregnancy is advisable (35). However, in the vast majority of these cases, the outcome is very poor, and termination of pregnancy constitutes reasonable advice, particularly if rupture of membranes occurred before 20 weeks of gestation with poor reaccumulation of amniotic fluid. Amnioinfusion has been proposed and attempted as a means of overcoming the problem of chest compression, with limited success thus far (36).

The management of a patient with impending premature delivery should include the following: evaluation of gestational age by dates and/or early ultrasound, fetal size and position, condition of the fetal membranes, amniotic fluid volume, and evidence of chorioamnionitis and other obstetric complications such as bleeding, toxemia, and so on. Vaginorectal cultures for the detection of group B streptococcal colonization and initiation of therapy with penicillin are also in order (37). If culture results return negative, penicillin can be discontinued. In all patients from 24 weeks of gestation who are not infected and for whom there is no maternal indication for immediate delivery, such as massive bleeding, and in whom the cervix is
dilated less than 5 cm, we propose tocolysis with magnesium sulfate and administration of betamethasone (29,38). Our experience over the years with the combination of tocolysis and betamethasone has been fully supported by the 1994 NIH Statement on Antenatal Use of Corticosteroids (39). For patients between 24 and 34 weeks’ gestation, we administer two doses of 12 mg of betamethasone, 24 hours apart. Beyond 34 weeks, we use steroids only when an amniocentesis indicates lung immaturity. Multiple pregnancies are offered similar therapy (40). We also monitor body temperature and changes in leukocyte count, keeping in mind the possible transient leukocytosis after the administration of betamethasone.

If a patient has fever or demonstrates other signs of chorioamnionitis, broad-spectrum antibiotics are initiated. In the presence of ruptured membranes, a number of obstetricians use a combination of ampicillin and erythromycin in an attempt to temporarily stop labor and administer steroids (41,42,43). This seems a reasonable approach, because between 30% and 50% of premature births are believed to be precipitated by common genital tract infections.

The successful management of the ELBW infant begins in the delivery room (Table 25-11). A well-organized and equipped delivery room and the presence of a competent team headed by an experienced neonatologist are essential ingredients to the proper reception of these very fragile newborns. The basic principle guiding successful management is directed toward prevention of any physiologic deviation from normality, such as hypothermia, acidosis, or hypoxia. At the same time, it is important that each intervention during the resuscitation process be carefully adapted to the size and to the needs of the tiny infant. Brisk maneuvers, excessive positive pressure with bagging, or inappropriate administration of drugs and fluids may induce permanent central nervous system (CNS) or lung injuries.

It seems particularly inappropriate when high-risk mothers are referred to a tertiary care center for specialized perinatal care, to have their premature newborn infants cared for in the delivery room and during the critical first hours of their lives by unsupervised and inexperienced in-training personnel. Major decisions, such as whether to initiate resuscitation and for how long, often need to be made in extremely short periods of time and under heavy pressure for infants at the limit of viability. This can be done only by experienced personnel (44).

In our center, the birth of an ELBW infant is always attended by a neonatologist in addition to the pediatric house staff and a trained delivery room nurse. Appropriate equipment is used according to the American Heart Association and American Academy of Pediatrics (AAP) guidelines for neonatal resuscitation, with particular emphasis on temperature control i.e., radiant heater set at maximum temperature and prewarmed blankets (45).

During the initial steps of stabilization, the condition of the infant is rapidly assessed. After drying, positioning on warm blankets, and suctioning, most ELBW infants require immediate initiation of positive pressure ventilation with a bag and mask. Resuscitation is initiated using an inspired oxygen concentration of 100%, which is rapidly reduced as the infant’s condition improves. Although concerns have recently been raised that such a practice may result in potential exposure to oxygen free radical species (46,47), the available data is limited to asphyxiated newborn infants at term, and there is currently insufficient evidence to support a change in practice. We found that, for ELBW infants, ventilation is more effective if performed at a higher ventilatory rate than for the term infant. We use anesthesia bags and ventilate at a rate of 60 to 80 breaths per minute, adjusting the pressure to provide adequate bilateral air entry and chest wall excursion. For extremely premature infants, intubation in the delivery room may rapidly follow.

With proper ventilation, in our experience, rarely will an infant require chest compressions or epinephrine. The prognosis of ELBW infants requiring this extent of resuscitation is very guarded, particularly if their birth weight is below 750 g. Fluid resuscitation is reserved only for those infants in which significant blood loss is suspected.

Even following optimal resuscitation, the Apgar scores of ELBW infants rarely exceed 6 or 7 in view of their decreased tone and reactivity, poor respiratory effort, and initially poor peripheral perfusion (48). The infant’s heart rate is thus the best measure of the effectiveness of resuscitation efforts.

The topic of delivery room management would not be properly covered without mentioning the ethical dilemmas faced by the neonatologist when parental and
medical opinions regarding resuscitation differ, or when an ELBW infant is severely asphyxiated and requires prolonged resuscitation. It is our view that reasonable parental opinions must be respected after full and honest discussion of the infant’s chances of meaningful survival.

Expert management in the delivery room and during the first hours after admission to the NICU is of paramount importance to prevent immediate and long-term complications in the ELBW infant. It is well established that the majority of cerebral injuries occur around the time of delivery or in the immediate postnatal period. Acute changes in cerebral blood flow may predispose the very fragile network of periventricular vessels to rupture. Hence, it is essential to handle these very fragile infants with extreme care, avoiding unnecessary disturbances, and preventing, rather than correcting physiological deviations in acid-base balance, blood gases, blood pressure, or body temperature. Also, overly aggressive ventilation either in the delivery room or in the NICU may predispose to significant acute or chronic pulmonary problems such as hyperinflation and loss of elasticity of the alveoli, pulmonary interstitial emphysema, pneumothorax, and eventually CLD.

The vast majority of our ELBW infants who require assisted ventilation are intubated in the NICU. Only in exceptional situations, when the infant does not respond to bag and mask ventilation, is intubation performed in the delivery room. We use the nasotracheal route, and an endotracheal tube (ETT) of 2.5 mm for infants with a birth weight less than 1000 g. We believe that it is important to use a low-caliber ET tube to avoid subglottic trauma, strictures, and eventually stenosis. We have never had to perform a tracheostomy in an infant, and stridor has been very rare among our patients.

If the newborn infant is in distress, an umbilical arterial catheter is inserted. The arterial line is used exclusively for blood sampling. We favor the “high” position of the tip of the catheter, just above the level of the diaphragm. After each blood sampling, the catheter is flushed with a heparinized solution of 0.45% saline. In infants weighing less than 750 g, we also insert an umbilical venous line. We use the venous line to infuse fluids, thus avoiding excessive handling and disturbance to the newborn infant during the first 24 to 48 hours of life. The tip of the catheter is positioned at the junction of the inferior vena cava and the right atrium, hence avoiding the liver. This is important, particularly when infusing hypertonic solutions, such as sodium bicarbonate or calcium gluconate.

Blood is analyzed for glucose, electrolytes, blood gases, hemoglobin, and leukocytes. An intravenous with 10% dextrose is initiated at a rate of 65 to 85 mL/kg/d and the infant is placed in a high humidity closed incubator. Serum glucose levels are closely monitored, and the concentration of dextrose administered is adjusted accordingly. In very tiny babies, when more than ten percent of the baby’s blood volume has been removed, we replace it with packed red cells. We use a single donor, collecting the blood in small packs, which may be used for up to several weeks (49). Donors are extensively screened for all viral illnesses. We encourage parents who wish to donate blood to their infants to do so, as long as they are of a compatible blood group and are free of viral and other infections. Very sick infants receive 1:1 nursing care until the condition is stabilized, at which point the ratio of nurse-to-baby becomes 1:2.

A percutaneous central venous catheter is usually inserted between the second and third day of life for intravenous alimentation (50,51). As portal of entry, we use the upper extremities of the infant, and we aim for the tip of the catheter to be in the superior vena cava, being careful to avoid intracardiac positioning, with its attendant risks of erosion into the pericardial space (52,53). In case of failure to properly position a central line, peripheral venous access is maintained using an extremity or scalp vein. Parenteral nutrition (TPN) is generally introduced within the first 24 hours of life. When the mother receives intravenous fluids during her labor, a baseline electrolytic profile of the newborn shortly after birth appears to be the proper way to follow subsequent changes. Electrolytes are repeated between 12 and 18 hours of age. During the first 72 hours, the body weight is recorded every 8 hours, and fluid intake is adjusted accordingly. The new incubators have incorporated scales, allowing recordings without excessive handling and disturbance of the newborn infant. One other major advantage of the new incubators is that they can provide a high level of humidity, hence substantially reducing the need for large volumes of fluid.

To establish prognostic criteria, it is important to obtain a cranial ultrasound in the first 24 hours of life (54,55). This ultrasound needs to be repeated at least one week later, or as often as necessary, depending on the pathology detected on admission or, if the infant’s condition has deteriorated, suggesting CNS involvement. It is also important, before discharge from the hospital, to repeat the cranial ultrasound to evaluate the presence or absence of periventricular leukomalacia (PVL) (56). Ideally, this last ultrasound should be performed at 35 to 36 weeks of postmenstrual age.

The vast majority of infants with a birth weight less than 1,000 g will need some form of respiratory assistance to survive. For vigorous infants, nasal continuous positive airway pressure (CPAP) or recently, nasal ventilation, is the preferred mode of support (57). Some controversy surrounds the timing and criteria for the initiation of assisted ventilation. Likewise, controversy also exists regarding whether these tiny infants should receive prophylactic exogenous surfactant in the delivery room (58,59,60,61,62). We do not systematically intubate infants born weighing less than 1,000 g, and we do not administer surfactant unless the infant requires assisted ventilation and a minimum FiO² of 0.30.

The introduction of exogenous surfactant therapy has reduced significantly the mortality of all newborns suffering from respiratory failure secondary to respiratory distress syndrome (RDS), but its impact has been particularly important among the most premature infants (63,64,65,66,67,68,69). Administration of surfactant in these very tiny infants requires extra care, as rapid changes in lung compliance may not only damage the lungs by creating overinflation and overdistention, but also may predispose to acute changes in ductal circulation which, in turn, could lead to both cerebral and/or pulmonary hemorrhage. With rapid improvement in oxygenation, persistent hyperoxia also may be detrimental to the eyes. Hence, the administration of surfactant should be performed by an experienced person, under close monitoring of ventilatory parameters, and with rapid reduction of peak inspiratory pressures (PIP) and oxygen concentrations. If necessary, a second dose of surfactant may be administered as soon as 6 hours after the first. However, in our experience, if the response to the second dose is not satisfactory, it is highly unlikely that the condition will improve with additional administration of surfactant. In our center, 64% of the babies improved rapidly, requiring only a single dose of surfactant. Natural surfactant preparations are nowadays practically the only ones used (70).

Mechanical ventilation has dramatically improved the survival of infants weighing less than 1,000 g. In the 1970s, very few infants born weighing less than 1,000 g survived. In the early 1980s, survival of infants weighing 500 to 750 g varied from 3% to 25%, and that of infants weighing 750 to 1,000 g ranged from 30% to 70% (71). In initiating mechanical ventilation, it is imperative that minimal effective settings be used (72). Studies have shown that hyperventilation and overinflation of the lungs increase the loss of surface active phospholipids (73). Also, overinflation predisposes to air leaks and particularly to pulmonary interstitial emphysema (PIE). The latter is a serious complication in the tiny infant, and is a relatively frequent one. It is probably related to structural immaturity of the lungs, particularly to the relative lack of elastic tissue, which normally increases progressively throughout gestation (74). Also, the interstitium is larger in the more immature infant as a result of poor alveolization. Although drainage of a pneumothorax may lead to rapid improvement, management of PIE is far more complicated. As lung compliance is reduced, there is a need for increased PIPs to maintain adequate ventilation. This results in increased barotrauma to the small airways. Chorioamnionitis has been reported as a risk factor predisposing to PIE (75). The highest incidence of PIE in tiny infants has been observed when intrauterine pneumonia complicates the RDS. To overcome the problems related to PIE, a number of strategies have been devised. These include acceptance of higher levels of partial pressure of CO₂ (PCO₂) and lower levels of potential of hydrogen (pH), reduction of the positive end-expiratory pressure (PEEP) to between 2 and 3 cm water (H₂O), selective intubation of the contralateral lung, positioning the infant on the affected side, increasing the expiratory time, and systemic corticosteroid therapy. The combination of the above strategies can occasionally produce quite spectacular recovery from this condition. However, high frequency oscillatory or JET ventilation is probably the most effective therapy (76,77).

A variety of ventilatory strategies have been promoted to maintain satisfactory ventilation and to reduce the risk of complications (78), such as high PIP-low rates, low PIP-high rates, variation in the I:E ratio, variations in the flow, permissive hypercapnia, tolerance of lower pH and, more recently, high-frequency oscillation and even ventilation via nasal prongs. In recent years, however, the general trend is to use the lowest possible PIP to achieve acceptable ventilation and oxygenation. Of course, the question is what is considered “acceptable”? Some neonatologists will tolerate a pH as low as 7.20 and a PCO₂ as high as 65 mm Hg. Most centers also aim for PaO₂ values between 50 and 70 mm Hg. Our own approach to the ventilation of tiny infants over the years has been to favor nasotracheal intubation with a 2.5-mm ET tube. Our PIPs rarely exceed 14 to 15 cm H₂O, and we set the PEEP at 5 cm H₂O, with initial rates of 65 to 75 per minute. We aim for arterial oxygen pressure (PaO₂) values between 45 and 50 mm Hg, which is enough to abolish production of lactic acid and, at the same time, remains relatively close to intrauterine values. Our pulse oximeters (79) are set to alarm at a lower limit of 80% and an upper limit of 93% for the first few weeks of life. We believe that this modest degree of oxygenation offers the advantage of reducing the need for administration of elevated oxygen concentrations, thus minimizing lung toxicity, and may help to avoid retinal damage. Our incidence of bronchopulmonary dysplasia (BPD) and ROP are shown in Table 25-9. We believe that by using the lowest possible PIP and initially, a relatively rapid respiratory rate, we reduce overdistention and barotrauma and minimize the risk of lung injury. Because, in RDS, there are compartments in the lung with relatively normal ventilation perfusion ratios and others with poor ventilation and adequate perfusion, it seems reasonable to attempt to improve ventilation of the poor ventilation perfusion (V-Q) compartment without overdistention of the normal V-Q compartment. Raising the ventilatory rate, which raises the mean airway pressure without changing the PIP, appears to accomplish this (80). We also have observed that with initially higher respiratory rates, the tiny infant very rapidly stops fighting the respirator, thus making the gas exchange smoother and possibly decreasing the incidence of air leaks. Relatively high respiratory rates also seem to be more physiologic for the very immature infant, as observed by Greenough and collaborators (81). For toilet of the airways, we use the Ballard closed suction circuit, thus avoiding disconnecting the infant from the ventilator (82). We suction sparingly during the first few days of life, when the volume of secretions is minimal. Analgesia/sedation is given for nonemergent intubation and for infants who remain agitated while on mechanical ventilation.

Most of our ventilated babies have their umbilical vessels cannulated for blood sampling and for fluid infusion.
As soon as the procedures of intubation and catheterization of the umbilical vessels are completed, we perform chest and abdominal radiography to assess the position of the ET tube and the umbilical catheters and, at the same time, to evaluate the severity of lung pathology. Fifteen minutes after the initiation of ventilation, we obtain an arterial blood gas and adjust the ventilatory parameters accordingly. We generally aim for a pH above 7.25 and a PCO₂ between 45 and 55 mm Hg, but when the PIPs are elevated or in the presence of PIE, we tolerate partial pressure of CO₂ (PCO₂) values up to 60 mm Hg as long as the pH is at least 7.20. Our ET tubes are sutured to the tape placed on the upper lip. We record the level at which sutures were placed on the ET tube, thus avoiding the need to repeat a chest radiograph to evaluate the tube position when reintubation is required. Actually, we take very few radiographs, and we rely extensively on clinical assessment, blood gases, and pulse oximetry. However, a chest radiograph should be taken if there are concerns about the position of the endotracheal tube or the development of any form of air leak.

Avery and associates (83) reported in 1987 that the incidence of BPD varied between neonatal units. The unit with the lowest incidence used CPAP much more frequently than did the other units. Epidemiologic data from 36 units in the Vermont-Oxford Trial Network also indicate large differences in the incidence of BPD, from 16% to 70% for infants weighing between 501 and 1,500 g (84). The incidence of BPD was lower in units allowing higher PCO₂ values. More evidence of the association of BPD and PCO₂ was provided by Kraybil and associates (85). More recently, Garland and associates (86) reported the highest incidence of BPD among infants with the lowest PCO₂ before the administration of surfactant.

The concept of permissive hypercapnia for patients requiring mechanical ventilation gives priority to the prevention or limitation of severe pulmonary hyperinflation over the maintenance of normal ventilation. The principle consists of allowing the PCO₂ to rise by minimizing ventilator pressures and tidal volume (87). Potential risks of high PCO₂ values include increased cerebral perfusion, increased retinal perfusion, increased pulmonary vascular resistance, and reduction of pH. Based on epidemiologic observations, it appears that respiratory acidosis, unlike metabolic acidosis, is not associated with poor neurologic outcomes. Vannucci and associates (88) demonstrated similar findings in animal studies involving rats.

Flow rate also can affect ventilation and increase airway injury. We generally use a flow rate between 3 and 5 L/min. Only when we need very high pressures, for instance, in the presence of pulmonary hemorrhage, do we allow the flow rate to exceed 5 L/min.

Several reports in the literature have expressed concern about potential side effects of low PCO₂ values (89). Graziani and associates (90) reported that, along with other factors, marked hypocarbia during the first 3 postnatal days was associated with increased risk of periventricular white matter injury in premature infants. The theoretical model of ischemic brain injury has been described by Wigglesworth and Pape (91). These authors hypothesize that cerebral blood flow could be decreased by several factors, including hypotension, hyperoxia, hypocarbia, and increased venous pressures. Concern also has been expressed in the literature about high-frequency ventilation, which may lead to low PCO₂ values as a result of effective alveolar ventilation (92). However, the data regarding the development of PVL among infants managed with these devices remain controversial. Most authors agree, however, that for hypocarbia to be dangerous for the brain, it has to reach levels below 30 mm Hg. Our policy is to avoid PCO₂ values below 40 mm Hg by first reducing PIP before reducing respiratory rates.

High frequency oscillatory ventilation (HFOV) has been used in recent years in an attempt to reduce the incidence of early ventilatory complications and to prevent bronchopulmonary dysplasia. Published reports are often contradictory and, so far, there is no clear evidence that HFOV offers an advantage over conventional ventilation (93,94). HFOV, however, offers an advantage when treating infants with PIE, pulmonary hypertension, or diaphragmatic hernias (95).

More recently, the effects of patient-triggered ventilation with volume-guarantee have been explored in the management of preterm infants (96). This novel technique calls for an automatic adjustment of the peak inspiratory pressure to ensure a minimum set mechanical tidal volume.

The timing of extubation of ELBW infants is very important, because they are prone to develop severe apnea, with the potential risk of cerebral injury. Nowadays, with early administration of surfactant and improvement in lung compliance, rapid extubation and placement on nasal CPAP or nasal ventilation is possible in the majority of ELBW infants. However, some extremely premature infants develop severe episodes of apnea and desaturation, requiring frequently reintubation. For this reason, for the tiniest infants, we often favor a more progressive weaning process by maintaining them for a few extra days at a very low PIP of 10 to 12 mmHg and rates of 15 to 25 per minute, although providing maximum intravenous and oral alimentation (97). When the infant is stronger and starting to gain weight, we administer caffeine and proceed directly to extubation. The infant then is placed on nasal CPAP. The CPAP is discontinued when, after periodic trials, the infant can maintain good oxygenation without significant apnea, bradycardia, and desaturations. If an infant on nasal CPAP shows signs of fatigue manifested by recurrent apnea and retention of CO₂, we try nasal ventilation prior to reintubation. In many circumstances, this approach provides the extra help that these tiny infants require to avoid reintubation (98,99).

By far, the major cardiovascular problem in ELBW infants is the presence of a patent ductus arteriosus (PDA). More than 50% of infants born weighing less than 1000 g will
have a PDA diagnosed during the first few days of life (100,101). The onset of clinical manifestations of the PDA is related to the timing of improvement of the infant’s respiratory status, which is associated with a decreasing pulmonary vascular resistance and a predominantly left-to-right shunt. The patency of the ductus arteriosus can be easily documented in the first hours of life, with the help of echocardiography. At this early stage of life, the shunt is either right-to-left or bidirectional, depending on the severity of the infant’s respiratory condition. In our center, the incidence of clinically significant PDA requiring therapy has been around 65% of all infants weighing less than 1,000 g. The left-to-right ductal shunting can be diagnosed as early as in the first day of life in infants with RDS who improved following surfactant therapy (102). An active precordium, with bounding pulses and visible carotid pulse, will often precede auscultation of a murmur. If left untreated, the infant may develop left-sided heart failure and pulmonary edema or hemorrhagic pulmonary edema, with significant deterioration of the respiratory status. Significant left-to-right ductal shunting may cause decreased peripheral perfusion and oxygen delivery. ELBW infants with significant PDA are at risk for IVH, necrotizing enterocolitis (NEC), renal failure, CLD, and metabolic acidosis (103). The size of the ductus arteriosus, and the ratio of the left atrium to aortic root can be easily measured by echocardiography (104). We consider as significant a PDA of diameter greater than 1.5 mm and/or a ratio of left atrium to aortic root greater than 1.3 for ELBW infants.

The ductus arteriosus of the premature infant is less responsive to the vasoconstrictive effect of oxygen and is less likely to close spontaneously than that of term infants, especially in infants with RDS. The classic management of PDA first involves medical and supportive measures, i.e., fluid restriction, diuresis, distending airway pressure, transfusion of packed red blood cells to keep the hematocrit above 0.4. If these measures fail to close the ductus arteriosus, pharmacologic closure is possible with a cyclooxygenase inhibitor, namely indomethacin (105,106,107). In ELBW infants, because closure of the ductus arteriosus is unlikely to occur spontaneously, the infant is at risk for the short- and long-term complications mentioned previously. Hence, therapy with indomethacin has become standard practice for the majority of these infants. Different protocols of treatment have been proposed. We treat most infants with 0.2 mg/kg of indomethacin every 12 hours for three doses. For infants developing clinical and/or biochemical signs of renal failure, we subsequently use 0.1 mg/kg per dose (108). A reduction in fluid intake is advisable prior to administration of indomethacin. In our experience about 80% of the treated infants will respond with a functional closure of the ductus. However, about 30% of these may reopen, in which case, further indomethacin is administered. If the ductus arteriosus fails to close after three courses of indomethacin, or if indomethacin cannot be administered because of significant renal dysfunction on previous treatment, then surgical ligation is necessary. In the last 8 years, 16.7% of ELBW infants required surgical ligation of the ductus after repeated failures of pharmacologic therapy. The more premature the infant, and the greater the postnatal age at the time of treatment, the greater the failure of indomethacin (109). A group particularly resistant to indomethacin therapy is ELBW infants with severe IUGR (26). Suggested hypotheses for this high failure rate of indomethacin therapy among this group of infants include chronic hypoxia, altered levels of prostaglandin, and altered number or sensitivity of their receptors (110).

Contraindications to indomethacin therapy include renal failure, active bleeding, and thrombocytopenia. The presence of IVH does not appear to be an absolute contraindication to the use of indomethacin. Recent studies indicate that there is no progression of the severity of IVH after administration of indomethacin for PDA closure (111). Until additional data are available, it is prudent in the presence of IVH to verify the platelet count prior to the administration of indomethacin and to eliminate any bleeding diathesis.

The optimal timing of administration of indomethacin has been a matter of debate. Although early or prophylactic use of indomethacin may result in a higher initial PDA closure rate and in a reduction in the incidence of severe IVH, it does not appear to ultimately reduce the need for surgical PDA ligation, does not confer any long-term respiratory advantage, and does not change the rate of survival without neurosensory impairment at 18 months (112,113).

Another cyclooxygenase inhibitor that seems very promising is ibuprofen (114). It has the theoretical advantage over indomethacin that it may increase the range of blood pressure at which cerebral blood flow is autoregulated, but has little effect on cerebral blood flow during normotension (115,116). In a Phase I study carried out in our center, we observed a dramatic response in 12 ELBW infants treated with ibuprofen within 3 hours of birth (117). All twelve infants had their PDA permanently closed after three doses of ibuprofen. We also observed a significant trend in the reduction of IVH and practically no side effects. These findings have since been confirmed by other groups of investigators. The efficacy of ibuprofen for the treatment of PDA appears to be equivalent to that of indomethacin, with a lower likelihood of oliguria (118). Furthermore, in a series of patients evaluated by near-infrared spectroscopy and Doppler ultrasonography, ibuprofen therapy did not result in a significant reduction in cerebral perfusion and oxygen availability (119). One reported side effect of ibuprofen, namely the development of transient severe pulmonary hypertension and hypoxemia in a few infants treated within six hours of birth, may still require further evaluation (120).