Why Does the Ductus Arteriosus Remain Open in Preterm Infants?

The ductus arteriosus closes in most term infants within hours after birth, but in the very premature infant the ductus frequently closes considerably later or fails to undergo spontaneous closure. This is in part due to the elevated sensitivity of the immature ductal tissue to the dilating effects of prostaglandins (PG) and the low sensitivity to the constrictive effects of oxygen. The ductus of the smaller preterm infant can remain open for days or weeks and in many cases, even when it may constrict initially, it can reopen later. This reopening is frequently associated with clinical deterioration induced by episodes of systemic infection or other events associated with a systemic inflammatory response, such as pneumonia or necrotizing enterocolitis. The incidence of a PDA is also increased in infants who are exposed to antenatal magnesium sulfate administered to the mother, while persistence of ductal patency after indomethacin therapy was observed in infants exposed to infection and inflammatory mediators before birth.

The incidence of a PDA is inversely related to gestational age and this is even more striking in preterm infants with respiratory failure. Some statistics indicate that more than 70% of preterm infants born before 28 weeks of gestation historically have been exposed to therapeutic interventions to close the PDA, with decreased efficacy of medical treatment at earlier gestational ages that had led to an increased need for surgical ligation. However, as a consequence of results of recent trials showing lack of prevention of bronchopulmonary dysplasia (BPD) with early PDA treatment, the substantial rate of spontaneous PDA closure, and the increasing use of noninvasive forms of respiratory support with less mechanical ventilation, fewer PDAs are being diagnosed and treated.

Despite multiple studies evaluating the effects of antenatal corticosteroids on neonatal outcomes, only a few evaluated their effect on the PDA. Evidence in experimental animals demonstrates a constrictive effect of glucocorticoids on the ductus and an increased responsiveness of the ductal muscle to oxygen. When premature infants were exposed to corticosteroids at least 24 hours before preterm delivery, the incidence of symptomatic PDA and PDA requiring treatment was significantly reduced. This effect was not observed when the steroids were administered to the mother less than 24 hours before delivery. There is also an association between lower cortisol levels in preterm infants during the first week after birth and a higher incidence of PDA. This may be explained by the fact that cortisol decreases the sensitivity of the ductal tissue to the dilatory effects of PGs.

With the introduction of exogenous surfactant therapy, the clinical presentation and incidence of patent ductus arteriosus in premature infants with respiratory distress syndrome (RDS) has been modified. Whereas surfactant itself has no effect on ductal contractility, the rapid improvement in the arterial partial pressure of oxygen (Pa o ₂ ) observed after surfactant administration can result in a rapid fall in pulmonary vascular resistance, producing an earlier clinical presentation of the ductus in preterm infants and in experimental animals. This may explain the results shown in a meta-analysis of several randomized trials that prophylactic administration of synthetic surfactant seems to increase the incidence of symptomatic PDA, as well as the possible association between surfactant administration and the development of pulmonary hemorrhage in infants with left-to-right ductal shunting. There is also a relationship between a large ductal diameter with significant shunting and pulmonary hemorrhage and early treatment (first 24 hours) with indomethacin in infants with a large PDA that shows significant reduction of pulmonary hemorrhage. Data from exogenous surfactant trials, when pooled, demonstrate an increased rate of symptomatic PDA and an increased risk of pulmonary hemorrhage, a serious complication associated with significant mortality and an increased risk of chronic respiratory morbidity. This is one clear pathway wherein a symptomatic PDA can result in lung injury and longer-term respiratory morbidity.

Systemic Consequences of a PDA

The consequences of a PDA in preterm infants depend on the size of the ductus and the magnitude of left-to-right shunting. This is determined by the difference in pressures between the systemic and pulmonary circulations. The left-to-right shunting results in increased pulmonary blood flow, volume overload of the left heart chambers, and decreased systemic flow and perfusion. While the left ventricle has the ability to increase its output, systemic blood flow distribution may be compromised by the decline in diastolic blood pressure and local vasoconstriction in different organs, resulting in decreased organ perfusion. This may explain many of the systemic manifestations of a significant PDA, such as renal dysfunction, poor gastrointestinal function, and increased risk of necrotizing enterocolitis. A significant PDA can also compromise cerebral blood flow, thus producing ischemia and contributing to the development of brain injury. The hemodynamic consequences of PDA are summarized in Fig. 7.1 .

Hemodynamic and respiratory consequences of patent ductus arteriosus. CNS, Central nervous system; GI, gastrointestinal; ICH, intracranial hemorrhage; NEC, necrotizing enterocolitis; PVL, periventricular leukomalacia.

Pulmonary Consequences of a PDA

PDA and BPD

Despite the strong experimental evidence that increased flow and pressure in the developing pulmonary vasculature can produce severe morphologic and functional alterations in the immature lung, there is still no conclusive evidence regarding the role of the PDA in the pathogenesis of BPD. The respiratory morbidity associated with a PDA is caused not only by the increase in pulmonary blood flow, edema, and pulmonary inflammation; many premature infants presenting with symptomatic PDA require mechanical ventilation and/or supplemental oxygen. The need for increased mean airway pressure and fraction of inspired oxygen (F io ₂ ) in the setting of increased left-to-right shunt may also be one of the factors in the causal pathway of BPD. Infants with PDA are exposed to multiple factors that increase the risk of lung injury, and these become important confounders in the reported association between PDA and increased risk for BPD. The negative effects of PDA on lung function were documented many years ago by Cotton et al. with the demonstration of a longer duration of mechanical ventilation among infants with symptomatic PDA compared with those without PDA. In addition, they reported a decreased duration of mechanical ventilation with early surgical PDA closure. The opposite was found when Clyman and colleagues reanalyzed the results of an earlier randomized, controlled clinical trial comparing early PDA ligation with expectant management and showed that infants who underwent early PDA ligation had a higher risk of developing BPD compared with the controls. A subsequent small trial of early pharmacologic PDA closure with indomethacin also revealed a reduction in BPD. It is important to note that these studies were conducted in the era before antenatal steroids and surfactant therapy.

From an epidemiologic standpoint, several studies since therapy with surfactant and antenatal steroids became available have found an increasing incidence of BPD with increased survival of extremely preterm infants, many of whom actually had mild or no initial RDS. In these populations multivariate logistic analysis showed an increased risk of BPD in infants with a symptomatic PDA and episodes of sepsis. The BPD risk was even greater when the PDA and sepsis occurred simultaneously, suggesting an interaction between these two events ( Fig. 7.2 ). Late reopening of a PDA was more frequent in infected infants, and failed PDA closure was more common when sepsis and PDA were temporally related. Moreover, infants with infection and those with PDA have higher levels of 6-keto-PGF _1α and elevated levels of tumor necrosis factor α (TNF-α). Elevated levels of 6-keto-PGF _1α in infected infants were found to be associated with an increased rate of late PDA and unresponsiveness to treatment with indomethacin. These data explain the association between infection and PDA outcome by increasing the risk of ductal reopening and closure failure.

Risk for development of bronchopulmonary dysplasia. Adjusted odds ratios and 95% confidence intervals obtained by multivariate logistic regression analysis show increased risk of developing bronchopulmonary dysplasia with the occurrence of patent ductus arteriosus (PDA) and when PDA is temporally related with sepsis. RDS, Respiratory distress syndrome.

Modified from Gonzalez A, Sosenko IR, Chandar J, et al. Influence of infection on patent ductus arteriosus and chronic lung disease in premature infants weighing 1000 grams or less. J Pediatr. 1996;128:470–478.

Despite the experimental and epidemiologic evidence supporting the role of the PDA in the pathogenesis of chronic lung injury, there are scarce data from more recent prospective clinical trials to confirm this association or demonstrate that treatment of the PDA has any effect on chronic lung injury.

Management of the PDA and Respiratory Outcome

Different strategies to close the PDA have been investigated in an attempt to reduce its adverse consequences. Closure of the PDA by surgical or pharmacologic means is generally associated with rapid improvement in lung mechanics, although it is uncertain whether this translates into improvement in clinical course. Several randomized trials conducted in the pre-surfactant era have compared early PDA closure with delayed treatment—that is, at a time when the PDA was deemed hemodynamically significant (by various definitions). In these trials, even when the time difference between early and late closure was relatively small, many studies demonstrated that early PDA closure was associated with decreased pulmonary morbidity. When Clyman conducted a meta-analysis of these earlier studies, results indicated that infants who received early PDA treatment were at lower risk for BPD and had a reduction in the duration of mechanical ventilation compared with those receiving delayed treatment. More recent trials, including the trial of Van Overmeire et al., did not find significant differences in respiratory outcome between infants with RDS and PDA who received indomethacin “early” (on day 3 ± 0.5) compared to “late” (on day 7 ± 1.7). However, in this study infants were relatively mature and half of the patients in the “late” treatment group had spontaneous PDA closure by day 9, suggesting that even the infants who were treated late were not exposed to the effects of the PDA for very long. Similarly, in a group of smaller and less mature infants, Sosenko and coworkers demonstrated lack of benefit from “early” PDA closure, compared with expectant management, with closure only when the PDA became hemodynamically significant. This study did not include infants with significant hemodynamic PDA at the time of enrollment.

Prophylactic indomethacin for the prevention of PDA has been extensively investigated. In most of these studies prophylactic indomethacin was started on the first day after birth, even before the onset of PDA symptoms. As expected, most of these reports found a significant reduction in the incidence of PDA and need for surgical ligation with indomethacin prophylaxis. Surprisingly, the most dramatic effect of prophylactic indomethacin was a reduction in the incidence of severe intracranial hemorrhage. However, in terms of pulmonary outcome, pooled data from two meta-analyses failed to demonstrate a decrease in long-term pulmonary morbidity when prophylactic indomethacin treatment (given before PDA symptoms) was compared with later treatment after the PDA became symptomatic. It is important to note that most infants randomly assigned to the control arm of these trials were exposed to a relatively short period of increased pulmonary blood flow because indomethacin was administered soon after the symptoms of PDA appeared. Another explanation for these findings is that the potential benefits of early PDA closure may have been negated by the detrimental effects of indomethacin on renal function and fluid retention, contributing to deterioration in lung function, particularly in those infants who did not have a PDA but still received prophylactic indomethacin.

Most recently published randomized controlled trials (RCTs) were not designed to assess the association between PDA and BPD as they are essentially early versus later treatment trials with little or no exposure of the untreated group to a prolonged systemic to pulmonary shunt. From these trials there is little evidence that a short-term exposure to a PDA leads to BPD. Newer studies have attempted to differentiate the effects of larger and more hemodynamically significant PDA by the use of scoring systems to delineate the magnitude of the left-to-right shunt. Larger PDA shunts were more predictive of subsequent death or chronic lung disease. Furthermore, the risk of BPD has been found to increase with the duration of exposure to a hemodynamically significant PDA.

When ibuprofen was compared with indomethacin for PDA closure in a meta-analysis of 6 randomized control trials, PDA treatment with intravenous ibuprofen after 24 hours was associated with an increased risk of BPD compared to treatment using indomethacin, RR 1.28, CI 1.03–1.60. However, no significant difference in BPD incidence was found in infants who received indomethacin versus placebo.

Paracetamol is a new option for medical treatment of the PDA with some apparent efficacy, especially for early treatment. It appears to have a less adverse vasoconstrictive effect and therefore may cause less fluid retention/impaired renal function. There is only one published placebo-controlled RCT of paracetamol for PDA treatment, which showed no associations with BPD either way.

In a recent extensive analysis of the literature, Benitz concluded that there was no clear evidence that routine medical or surgical closure of the ductus was beneficial in preterm infants. However, this conclusion is not entirely supported by the results of the review that showed a significant decrease in death or chronic lung disease in infants who received early treatment compared with those treated late for symptomatic PDA. Another important issue regarding this systematic review is the variation in the studies included. Many were performed more than 20 years ago, with poorly defined diagnosis and enrolment criteria, in more mature infants than are treated today, and with high rates of open-label treatment. Other large retrospective studies of different treatment approaches to PDA in extremely low-birth-weight infants have shown higher mortality in the conservative group and more BPD in the surgical closure group versus pharmacologic treatment that was protective for death/BPD at 36 weeks, although not all cohort trials have found these results.

The main argument of those proposing a conservative approach to interventions to close the PDA is the fact that many will close spontaneously, and waiting for spontaneous closure prevents the unnecessary use of drugs or surgery, both of which are associated with significant complications. However, the rate of spontaneous PDA closure is dependent on gestational age and is much lower in an infant with a birth weight less than 1000 g who has severe lung disease and is already at higher risk of severe BPD. There is no experimental or clinical evidence that the persistence of an open ductus (especially one causing hemodynamic compromise) confers any benefit in the clinical course of premature infants. Therefore the risk/benefit ratio must be measured in each individual case (i.e., the possible negative consequences of a persistent open ductus vs. the complications associated with the interventions to close it). Without definitive data, the approach to PDA treatment must be individualized to each infant, taking into account the gestational age, clinical course, size, and hemodynamic consequences of the ductal shunt and the potential side effects of the interventions to close the PDA ( Fig. 7.3 ). The relatively low and unpredictable efficacy of the available medical treatments compounds the problem and provides an area in which research to predict which infants would benefit from therapy, as well as the optimal dosing and timing of treatment, would be useful.

Clinical considerations in the management of patent ductus arteriosus (PDA). BPD, Bronchopulmonary dysplasia; COX, cyclooxygenase.

Most infants who have a PDA and develop BPD are extremely premature, but surprisingly few studies have focused an analysis on this population. When Mahony et al. compared the effects of early versus late medical treatment of PDA, early treatment did not reduce oxygen dependency in the overall population. However, stratified birth weight analysis showed early treatment was associated with a significantly shorter duration of oxygen need in infants with birth weights less than 1000 g. This finding was not seen in the more recent study of Kluckow and coworkers, who were unable to demonstrate a decrease in BPD with early indomethacin in infants less than 29 weeks of GA, or in the study by Sosenko et al. in which early administration of ibuprofen in infants with birth weights less than 800 g actually resulted in an increased need for oxygen at concentrations greater than 30% at 36 weeks postmenstrual age compared with the control group. One of the major obstacles in deciding whether early treatment is superior to a conservative approach is that a large proportion of infants assigned to the conservative arms (placebo) in the various trials eventually received open-label treatment when the PDA became significant. In these cases, open-label treatment reduced the exposure of the infants to the negative effects of the PDA. Thus when the treatment strategy of one arm of the trial becomes the rescue therapy for the other, it poses a significant challenge to those designing clinical trials to balance ethical and scientific standards.

Studies that evaluated the effect of short versus prolonged indomethacin therapy did not show consistent effects on PDA closure or on respiratory outcome, but the strategies that have led to more effective PDA closure have also resulted in lower respiratory morbidity.

When pharmacologic treatment is contraindicated or fails, surgical ligation of the PDA is generally used as a second alternative. As with medical treatment, the potential benefits of surgical closure must be weighed against the potential side effects of the surgical procedure. The effect of surgical PDA closure on lung mechanics is an immediate improvement, with a rapid increase in compliance, most likely due to the sudden reduction in pulmonary blood flow, blood volume, and interstitial lung fluid. Despite these positive hemodynamic and pulmonary changes from surgical PDA ligation, the clinical condition of a significant number of infants deteriorates after surgery. This deterioration, manifested by higher inspired oxygen requirement and arterial hypotension, occurs mostly in smaller infants who had a prolonged left to right shunt via the PDA. The mechanisms for this may be related to surgical trauma, to left ventricular dysfunction caused by an immediate increase in afterload, or a sudden reduction in pulmonary blood flow after ductal closure. The early use of milrinone in this population appears to improve cardiovascular function. Studies of surgical ligation in a baboon model have demonstrated a significant upregulation of genes involved with pulmonary inflammation (cyclooxygenase 2-2, TNF-α, and CD14), which may explain deterioration in some infants, or why postligation pulmonary mechanics fail to improve.

Because no trials have compared surgical closure of a hemodynamically significant PDA with continued medical management when this has been ineffective, the benefits or side effects of surgical ligation on respiratory function remain unclear. However, recent reports have suggested significantly worse neurologic and pulmonary outcomes in infants who had PDA ligation versus those who had successful pharmacologic closure or needed no treatment. Several other complications have been associated with PDA ligation, all of which can be associated with an increase in respiratory morbidity, including compromised cerebral oxygen supply, diaphragmatic paralysis owing to phrenic nerve injury, chylothorax, scoliosis, and vocal cord paralysis. As previously mentioned, Clyman et al. reanalyzed data from a previous RCT of early PDA ligation during the first days of life and found a threefold increase in risk of developing BPD with early PDA ligation compared with the controls. Most recently Janz-Robinson et al. found that both medical and surgical PDA treatment were independent risk factors associated with poor neurodevelopmental outcomes in 29-week-old preterm infants.

In another interesting finding, Schmidt and coworkers showed that caffeine, administered to decrease apnea and favor successful weaning from mechanical ventilation in premature infants, lowered the incidence of PDA, decreased the need for PDA ligation, and also decreased the incidence of BPD compared with infants who received placebo. The mechanisms for these beneficial effects are not clear but besides favoring weaning from mechanical ventilation by stimulating central respiratory activity, caffeine may function through its weak diuretic effect or through its antiinflammatory properties.

The persistence of a PDA in critically ill preterm infants can be associated with acute deleterious effects and an increased risk of BPD. Although the effects of the ductus on lung development and function are likely related to the size of the ductus and duration and magnitude of left-to-right shunting, more data are needed to define the impact of a more conservative approach toward the PDA on respiratory, cardiovascular, and neurologic outcomes. These trials should clearly separate treatment from nontreatment arms by increasing drug efficacy, choosing patients likely to have a high treatment effect, and avoiding open-label treatment in the placebo group. Until further evidence becomes available, the decision on management of PDA must balance the consequences of ductal patency that are closely related to its hemodynamic significance versus the side effects of the interventions to close it (see Fig. 7.3 ).