The tendency of the ductus arteriosus (DA) to remain patent in preterm infants is thought to be related to developmental immaturity of ductal anatomy,¹ as well as altered prostaglandin responsiveness² and metabolism (see also Chapter 16).³ Not all PDAs are pathological, however, even in the preterm population. It is important to consider the PDA in the context of the clinical picture and associated hemodynamic state of the patient. For example, a PDA may be considered supportive in the setting of critically low pulmonary blood flow (either due to right ventricular outflow tract obstruction or severe pulmonary hypertension), left ventricular (LV) dysfunction, or left heart obstruction (e.g., coarctation of the aorta). A PDA is considered hemodynamically significant, and potentially pathological, when there are clinical and/or echocardiographic markers of a moderate or large volume left-to-right (systemic-to-pulmonary) shunt.

Flow through the DA is determined by the Poiseuille law (Q = ΔPπr⁴/8Lμ), where Q denotes flow, ΔP is the pressure gradient across the ductus, L is the ductal length, and μ is the viscosity (see Chapter 16). This has important implications for understanding factors that modulate ductal flow. The pressure difference in the systemic and pulmonary circulations is an important determinant of transductal shunt, particularly during the postnatal transition. The high systemic vascular resistance (SVR) and declining pulmonary vascular resistance (PVR) state of the transitioning preterm neonate creates a unique physiological vulnerability to high-volume left-to-right ductal shunting. The inherent diastolic “dysfunction” of the preterm myocardium and poor adaptive ability in the face of adverse loading conditions further compounds this risk. Physiologically, the negative consequences of an hsPDA can be considered complications relating to one or more among pulmonary over-circulation, systemic hypoperfusion, and/or hypoperfusion-reperfusion, which arise as a result of a significant left-to-right transductal shunt. Retrograde diastolic blood flow (ductal “steal”) diverts blood flow from post-ductal organs, such as the kidneys and gut, resulting in hypoperfusion. Conversely, the increased left ventricular output (LVO) secondary to left-to-right ductal shunting increases blood flow to pre-ductal organs, such as the brain. The latter is a potential mechanism for both peri-/intraventricular hemorrhage (P/IVH) associated with reperfusion, and white matter injury.⁴ As is evident from the Poiseuille equation, blood viscosity, and specifically red cell mass, blood platelet levels, and intravascular volume status, also have the potential to modulate flow across the PDA and should not be disregarded when considering factors contributing to the ductal flow pattern. Although PDA size is frequently ascertained as part of a comprehensive hemodynamic assessment, ductal morphology is less commonly evaluated. Several different anatomical subtypes of PDA exist,⁵ which may contribute to the limitations of using PDA size as a single marker from which to estimate shunt volume. The several factors regulating PDA flow highlight the importance of a comprehensive hemodynamic assessment when appraising the ductus and embarking on treatment-related decisions. The concept of hsPDA involves complex interactions between intrinsic, patient-related conditions, the anatomy of the PDA itself, and the medical therapies used as part of newborn intensive care practice. In summary, relevant considerations include:

Although the relative contribution of the preterm patent ductus to important neonatal morbidities has been studied extensively and evidence of association is strong, the role and merits of treatment remain controversial. The argument that the PDA is an innocent bystander with high likelihood of spontaneous closure, rather than a pathologic condition, is becoming increasingly prevalent in neonatology.⁶ As a result, there is a secular trend toward a conservative or non-intervention approach to PDA. Opting for non-treatment and awaiting spontaneous closure implies acceptance of any short- and long-term consequences of a PDA, which are in part related to the timing of intervention. The adverse effects of hsPDA relate to the duration and magnitude of exposure to a left-to-right shunt, inherent circulatory adaptive mechanisms,⁴ and interplay of important clinical factors (e.g., gestational age, comorbidities, chronological age, end-organ function). Prolonged patency is associated with numerous adverse outcomes, including prolongation of assisted ventilation and higher rates of BPD, pulmonary hemorrhage, P/IVH, periventricular leukomalacia (PVL), renal impairment, necrotizing enterocolitis (NEC), systemic hypotension, and death.^7,32 The key issue is whether intervention at a particular time point prevents some or all of these complications with minimal side-effects related to treatment. Importantly, for each individual infant, there will be a different risk/benefit equation.

In most newborn infants, even in the first postnatal hours, the ductal shunt is completely left-to-right or bidirectional with a dominant left-to-right component, illustrating that pulmonary pressures are usually sub-systemic shortly after birth. Using superior vena cava (SVC) flow as a surrogate measure of systemic blood flow, a negative association has been observed between duct diameter and SVC flow at 5 hours of age, but this association was not significant in subsequent studies at 12, 24, and 48 hours.⁸ The association between the early low systemic blood flow and development of P/IVH and later necrotizing enterocolitis (NEC) suggests a possible mechanism by which PDA shunting contributes to the pathophysiology of these conditions.⁸ There is also mounting evidence to suggest a PDA may cause pulmonary hemorrhage in preterm neonates because of overload of the pulmonary circulation in the presence of a low resistance pulmonary vasculature and that early ductal treatment may prevent this.^9–11

A pathological PDA causes high left-to-right shunt volumes, which may flood the lungs and cause pulmonary edema. Pulmonary edema reduces lung compliance, resulting in increased ventilator and oxygen requirements. All these factors together might contribute to the development of BPD, known to be associated with the persistence of an hsPDA. Each week of exposure to a hemodynamically significant DA represented an added risk for BPD (odds ratio 1.7).¹² A higher incidence of BPD has also been observed in extremely preterm infants receiving conservative management of PDA, compared with infants without PDA.¹³ There is emerging but not unequivocal¹⁴ evidence that a tolerant approach to PDA may be associated with a higher incidence of BPD, particularly if treatment is delayed until after the first postnatal week.^13,15,16 Use of multiparameter scoring systems may also be predictive of future BPD and death – a high PDA severity score on day 2 is associated with these outcomes.¹⁷

A PDA may cause hemodynamic disturbances, resulting in “steal” of blood from the systemic circulation, including the mesenteric arteries, with consequences of decreased oxygen delivery to the gut and the potential for tissue injury and NEC. Even a low-volume PDA shunt can reduce mesenteric artery flow and decrease the expected postprandial increase in blood flow.¹⁸ Because reduced intestinal blood flow is a contributor to the development of NEC, hsPDA may be a causative factor for NEC.¹⁹ In a study involving a relatively large number of neonates, presence of a PDA was an independent risk factor for the development of NEC in VLBW infants.²⁰

There is ample evidence on the basis of Doppler and near-infrared spectroscopy (NIRS) studies to suggest that cerebral blood flow is reduced in the presence of a PDA.¹⁸ In a recent study using NIRS, cerebral tissue oxygen saturation was lowest in a group of newborns just prior to the surgical closure of PDA. Magnetic resonance imaging (MRI)–measured global and regional cerebral (and cerebellar) volumes were lower in the subgroup of infants that met criteria for surgical ligation compared with patients treated medically and those without a PDA. The surgical group also had a lower cerebellar volume compared with other groups. The authors speculated that prolonged exposure to left-to-right shunting, based on the amount of time elapsed between the diagnosis of PDA and actual surgical closure, was contributary.²¹ Although hsPDA has an effect on cerebral hemodynamics, whether it is causative for P/IVH is a question that remains unanswered. Cerebral autoregulation is likely to play some role, particularly in immature infants, in protecting against P/IVH (see Chapter 7 for pathophysiology of P/IVH). Intact autoregulation is variable in immature infants,²² and one of the risk factors for impaired autoregulation may be a PDA-attributable reduction in cerebral blood flow. Early (in particular prophylactic) treatment with indomethacin results in both closure of the PDA and decreased risk of severe P/IVH. However, many clinicians are not convinced that prophylactic or early targeted treatment of the PDA is helpful because of the lack of demonstration of improved neuro developmental outcomes.⁴² PDA is also a risk factor for development of PVL.²³ Finally, PDA is associated with a higher mortality rate.^7,32 In a retrospective study, after adjustment for perinatal factors, level of maturity, disease severity, and morbid pathologies, the hazard risk for death in neonates with a PDA was eightfold higher than in those with a closed ductus.²⁴ Exclusion of patients who died during the first 2 weeks or inclusion of those who underwent ductal ligation did not alter the findings. In neonates born prior to 28 weeks of gestation a PDA diameter ≥1.5 mm on postnatal day 3 was associated with greater odds of mortality.²⁵

What about spontaneous closure of a PDA – is it possible to do nothing and just wait?

Due to reduced spontaneous ductal closure rates coupled with significant pulmonary-systemic pressure differences and immature cardiovascular adaptive responses, extremely preterm infants (<28 weeks’ gestation) are at higher risk of complications from hsPDA. These risks exist both in the early postnatal period, with susceptibility to pulmonary flooding and systemic hypoperfusion and their consequences, as well as on a more chronic basis due to effects of prolonged left-to-right shunting on the developing heart, brain, kidneys, intestines, and lungs. A PDA in a relatively mature preterm neonate is of less concern, as the left-to-right shunt appears to be better tolerated and the cardiovascular system and cerebral autoregulation protective mechanisms are more developed. Despite this, previous systematic meta-analyses regarding PDA treatment include many trials more than 20 years old and those that focused on larger, more mature infants of up to 33–34 weeks’ gestation, where spontaneous closure in the first few days is almost inevitable.²⁶ Inclusion of both extremely preterm and relatively mature neonates in the same studies makes it difficult to understand the efficacy of treatment versus the effects of spontaneous closure. The protagonists of no treatment argue on the basis of these studies, despite inclusion of a wide mix of gestational age groups, poor diagnostic criteria for PDA, including limited and inconsistent assessment of hemodynamic significance, high rates of open-label treatment in control groups, small sample sizes, selection bias, and/or lack of objective enrolment criteria at randomization. Importantly, spontaneous closure in the placebo control arm of the randomized controlled trials, as well as high levels of open-label treatment in both trial arms (from 30 to 70%),²⁷ make the interpretation of outcomes in many of the PDA trials difficult at best. In fact, much of the efficacy of our current treatment drugs may be ascribed more to high spontaneous closure rates, particularly in more mature infants, rather than the treatments themselves. The real natural course of PDA in treatment-naïve extremely preterm neonates and the true efficacy of the medications used for treatment are not well understood, although literature on outcomes of non-intervention has evolved in more recent years.

Interest in the role of non-treatment or conservative management of the PDA has prompted a number of studies attempting to address these questions. In a study of the natural evolution of PDA in extremely-low-birth-weight (ELBW) neonates,⁵ the authors observed a 73% spontaneous DA closure rate in newborns born at less than 28 weeks. However, deaths (both early and late), undiagnosed probable PDAs, and infants discharged home with persistent PDAs were excluded from the study. In the end, 41% of potentially eligible neonates were excluded, many with morbidity and mortality potentially attributable to or contributed to by the PDA, including pulmonary hemorrhage, severe P/IVH, and hypoxic respiratory failure. This significantly impacts upon our ability to draw conclusions about the potential risks of non-treatment based on this study, noting also that a high incidence of pulmonary hemorrhage (25%) was observed.²⁸ Similarly, Koch et al. reported a 34% permanent closure rate of PDA in ELBW neonates.²⁹ Of note, around one-quarter of potentially eligible infants were omitted as a result of either death or provision of comfort care, and there was no use of cardiac ultrasound (US) to adjudicate hemodynamic significance. Despite these shortcomings, the observation that for each week of increase in gestational age above 23 weeks, the odds of spontaneous PDA closure increased by a ratio of 1.5 is noteworthy. This is consistent with the known direct relationship between gestational age, birth weight, and persistence of the PDA.²⁴ Spontaneous closure rates of 21–31% among infants with 23–27 weeks of gestation have been reported historically, with lowest rates in the most immature infants.^30,31 It is important to note that these earlier studies were undertaken prior to the widespread adoption of non-invasive ventilation for extremely preterm newborns and preceding the higher survival rates currently achieved at <25 weeks of gestation.

Two larger observational studies of minimal/no treatment of the PDA have been published from the same group. The first study¹⁴ compares three different PDA management approaches in 138 VLBW infants. Infants received either symptomatic, early targeted (during the first 48 hours), or conservative treatment. The authors found no short-term differences between the groups and a decreased rate of BPD in the conservative treatment group. The second, more contemporaneous study³² is a retrospective cohort study in two European units that enrolled 297 VLBW infants, of whom 280 received conservative PDA management. The authors documented a median time to PDA closure of 71 days for infants born at <26 weeks, compared with 8 days at 28–29 weeks and just 6 days for infants born at or above 30 weeks.³² Despite closure rates at hospital discharge of 85% for the overall cohort, spontaneous closure rates <750 g were particularly low, with a substantial proportion of these infants experiencing a prolonged period of exposure to ductal shunting.³² Detailed assessment of PDA hemodynamic significance was not provided; however, it can be assumed that a subset of patients, perhaps 30–50%,³³ would have had a moderate or large volume left-to-right shunt. In addition, of 26 infants who died, 16 had a cause of death potentially related to PDA. In 2022, among a large cohort of infants discharged home with a PDA after preterm birth, lower rates of spontaneous closure than previously observed (47–58%) were reported in the first 12–18 months after birth.³⁴ There were also three infant deaths in this group, of whom two had documented evidence of pulmonary hypertension, which was progressive.³⁴ A greater likelihood of BPD and death has also been reported elsewhere with non-treatment,^35,36 with an increase in BPD rates of 31% in one study in an era of non-intervention.³⁷ Pulmonary vascular remodeling as a result of chronic pulmonary over-circulation has been postulated as a mechanism for some of these observations, as well as a potential driver of preterm ductal closure.³⁸ Elevations in pulmonary arterial pressure are initially due to increased pulmonary blood flow (PBF) in the setting of hsPDA, which chronically may relate to pulmonary vascular remodeling and altered vasoreactivity.^39,40 Increases in PVR may eventually be substantial enough to lead to PDA shunt reversal in some patients.^39,40 These studies highlight that non-intervention of PDA is not necessarily benign, even if early morbidities such as P/IVH and pulmonary hemorrhage are avoided. They also serve to emphasize the importance of considering individual patient factors, such as gestational age and PDA hemodynamic significance, in clinical decision-making. Notwithstanding the limitations of the literature, it is clear that spontaneous closure rates are lowest among the smallest, most preterm infants, who are also the most vulnerable to PDA-related morbidities.

The role of conservative management of PDA is an important area of research, although there are no data focusing on the impact in the most extremely preterm infants with prolonged exposure to moderate-high volume shunts. A systematic review and meta-analysis on outcomes of conservative management of PDA, which encompassed 12 cohort studies and 4 randomized controlled trials (RCTs), was recently published.⁴¹ Findings of cohort studies included a higher risk for mortality (risk ratio 1.34) and a lower risk of BPD, P/IVH, NEC, and retinopathy of prematurity with conservative management.⁴¹ Reductions in prematurity complications may be contributed to by survival bias, however. In a subgroup analysis of two cohort studies where echocardiographic criteria for PDA were used, a lower risk of BPD in the group receiving conservative treatment, with no difference in mortality or other morbidities between the two groups, was noted. Meta-analyses of the RCTs showed no difference in outcome with conservative management when compared with intervention (pharmacological or surgical).⁴¹ This is consistent with findings from a network meta-analysis of all major treatment options for PDA, including placebo or non-treatment,⁴² although relatively high rates of open-label treatment in the “no treatment” arms were observed in included studies. Two recent trials documented lower rates of open-label treatment and may offer a more accurate assessment of the effects of true conservative PDA management. One was a small pilot trial of 72 infants receiving either intravenous NSAID treatment or placebo in the first 72 hours after birth, with no difference in secondary outcomes.⁴³ The second randomized 142 infants to oral ibuprofen or placebo between postnatal days 6 and 14⁴⁴ and, although no differences were identified in clinical outcomes, the authors noted low efficacy of NSAIDs, particularly in the most preterm infants. Although these are small studies, they provide a starting point for design of larger trials investigating conservative management without a requirement for high rates of open-label NSAID treatments as “rescue”.

Lastly, it is worth drawing a distinction between non-treatment and “conservative treatment” that involves shunt modulation strategies to attempt to reduce ductal flow or treatments to improve heart failure symptoms, without the use of medications or surgery for PDA closure. These include fluid restriction, use of diuretics, positive end-expiratory pressure (PEEP), postnatal corticosteroids, targeted hematocrit, hemoglobin level or platelet thresholds, and increased caloric intake. There are limited data available on the efficacy of these supportive care measures, and documentation of how strategies such as these are utilized in clinical trials of pharmacological PDA treatment is variable. In general, judicious fluid management is suggested, in part based on evidence that excessive fluid intake in the first week after birth is associated with an increased incidence of both hsPDA and NEC.⁴⁵ There is not strong evidence for fluid restriction, however, and an overly restricted fluid intake has been associated with a greater risk of acute kidney injury⁴⁴ as well as concerns about inadequate nutritional intake. In addition, theoretically, reducing overall circulating fluid volume without any effect on the proportion of cardiac output that is directed toward the lungs may risk exacerbating post-ductal hypoperfusion from hsPDA. Cerebral hypoperfusion is an additional risk, with a small prospective multicenter study demonstrating a significant reduction in SVC flow with fluid restriction for PDA.⁴⁶

Furosemide, a loop diuretic, is commonly used for treatment of congestive cardiac failure, including in neonates. The role of frusemide in modulating ductal shunt has not been well studied due to historical concerns, including prostaglandin-mediated dilatation of ductus arteriosus vessels in an animal model⁴⁷ and an increased incidence of PDA in an RCT where the drug was used.⁴⁸ There is also considerable overlap in symptomatology between longstanding hsPDA with pulmonary congestion, and BPD, which makes the role of diuretics in this setting difficult to study. More evidence is needed, particularly in infants with pulmonary venous congestion due to diastolic dysfunction, to ascertain the role of diuretics in supportive care of infants with PDA. Appraisal of diastolic function has therefore been suggested as a means of more carefully selecting patients for a trial of diuretic therapy from a research perspective.⁴³

Increasing PVR through use of PEEP, generally through use of continuous positive airway pressure (CPAP) for respiratory support, has been implemented but has not been shown to produce changes in lung function (from reduced pulmonary congestion),⁴⁹ although there is a suggestion ductal flow may be reduced.⁵⁰ Physiologically, higher hematocrit, and hence greater blood viscosity, has the potential to limit shunt volume. However, clinical trials of targeted hematocrit to date – either by delayed cord clamping or liberal blood transfusion thresholds – have not shown a reduction in NSAID exposure or surgical ligation for PDA treatment.^51,52 A more liberal approach to platelet transfusion has similarly not translated into meaningful clinical benefit in the trial setting⁵³ and is not presently routinely recommended.