Pathogenesis

In the normal term infant, the ductus arteriosus constricts soon after birth, stimulated by the rapid postnatal increase in arterial oxygen tension. The muscularis of the ductus is uniquely responsive to oxygen, reacting to an increase in ambient oxygen with sustained contraction. The mechanisms of this response are not fully understood, but the cytochrome P450 system (CYP450), endothelin-1 (ET-1), and intracellular oxidation-reduction (redox) balance appear to be important.¹⁷ Glucocorticoids may have an essential role in maturation of the ductal oxygen-sensing apparatus, as lack of antenatal exposure to glucocorticoids is associated with both decreased expression of genes for calcium and potassium channels implicated in the oxygen response, as well as with increased risk of persistent ductal patency. This direct response is augmented by a rapid decline in circulating levels of PGE₂, a potent relaxant for ductal smooth muscle that has a major role in keeping the ductus arteriosus open in the fetus. The placenta is the primary source of fetal PGE₂, so circulating levels fall precipitously upon separation from the placenta. Maximal effects of PGE₂ withdrawal on ductal closure seem to require antenatal priming of the ductal muscle by the rising levels of PGE₂ normally seen late in gestation. Dynamic functional closure usually is complete within the first 4 days after birth, but anatomic obliteration is not achieved until after 1 week of age.

In otherwise normal infants born at or near term, it is unusual for the ductus arteriosus to fail to close within the first 4 days. With decreasing gestational age at birth, both the age at spontaneous closure and the proportion of infants with PDA at any postnatal age progressively increase. Before interventions to close the PDA in preterm infants were widely adopted, ductal closure nearly always occurred spontaneously if given sufficient time, sometimes at 4 to 6 months of age.⁴⁹ Those observations date from an era when VLBW infants rarely and ELBW infants essentially never survived, long before adoption of antenatal steroid treatment and availability of exogenous surfactant. More current data indicate that the ductus arteriosus is very likely to close without intervention in babies born at greater than 28 weeks’ gestation⁴⁰ who weigh greater than 1000 g at birth,⁴⁷ or who do not have RDS.⁵³ In nearly all reported series of less mature, smaller infants with RDS, PDA was eventually treated to achieve ductal closure, typically late in or after the first week after birth. Consequently, it is impossible to know when or how frequently the ductus might ultimately close spontaneously in such infants, or what risk factors might predict failure of that process. Nonetheless, data from the placebo arms of clinical trials indicate that many such infants never develop signs of a hemodynamically significant ductal shunt. For example, the ductus arteriosus closed by 3 days of age in 60% of subjects in a trial of ibuprofen prophylaxis in infants less than 31 weeks’ gestation.⁶³ In a trial of indomethacin prophylaxis, 50% of infants born at 500 to 999 g never developed signs of a hemodynamically significant PDA.⁵⁶ Among 26- to 31-week gestation infants receiving ventilatory assistance in whom PDA was confirmed by echocardiography on day 3, the ductus arteriosus closed by 9 days of age in 78% without treatment.⁶⁴ Therefore, spontaneous closure in these high-risk groups is not exceptional.

The reasons for delay in or failure of ductal closure in preterm infants have not been fully elucidated. Birth before term may simply come before the ductus is prepared, functionally or structurally, to respond to signals that normally induce closure. Genes essential to ductal responses, including K⁺ and Ca²⁺ channels⁶¹ involved in muscle contraction and phosphodiesterases⁴² that degrade cyclic guanosine monophosphate (cGMP) and cyclic adenosine monophosphate (cAMP) mediators of the vasodilator responses to nitric oxide (NO) and PGE₂, are developmentally upregulated in the ductus of the fetus approaching term, for example. Lack of preconditioning by PGE₂ in late gestation may make the ductus arteriosus less sensitive to withdrawal of that vasorelaxant. Inadequate glucocorticoid production, common in extremely preterm infants, may compromise oxygen-sensing pathways, particularly in infants not exposed to maternal antenatal steroids.¹² Cytokines produced by a fetal inflammatory response, a frequent correlate of preterm birth, may directly impede ductal responses.³⁹ These potential mechanisms correspond to the clinical observations that PDA is strongly associated with extreme immaturity, lack of exposure to antenatal steroids, and signs of fetal inflammation or infection.

Pathophysiology

Before birth, the pulmonary vascular resistance is high and more than 90% of right ventricular output flows right-to-left through the ductus arteriosus. With the first breath, lung inflation and oxygenation produce a rapid decline in pulmonary vascular resistance, a decrease in pulmonary arterial pressures, and a large increase in pulmonary blood flow. Removal of the low-resistance placenta from the systemic circulation increases systemic arterial pressures, and flow through the ductus arteriosus quickly reverses direction to left-to-right. Pulmonary blood flow substantially exceeds systemic cardiac output until the ductus arteriosus closes. If ductal closure occurs in the first few days, this imbalance in distribution of the total cardiac output is well tolerated. If the ductus arteriosus remains open, however, the normal decline in pulmonary vascular resistance over the first several days and weeks results in progressively increasing left-to-right shunting through the PDA, with excessive pulmonary blood flow, increased volume work for the left heart, and the potential for systemic ischemia.

Infants can tolerate large left-to-right shunts with high pulmonary-to-systemic blood flow ratios (Q_P : Q_S) without developing pulmonary edema. In preterm infants, surfactant deficiency, low serum oncotic pressures, and compromised capillary integrity (e.g., that which accompanies bacteremia) lower the threshold for development of pulmonary edema. Accumulation of alveolar and interstitial fluid increases the alveolar-arterial oxygen gradient and reduces lung compliance. A large left-to-right shunt increases volume work for the left heart, leading to atrial and ventricular enlargement and myocardial dysfunction. This may produce electrocardiographic and serum biomarker signs of myocardial ischemia, including ST segment depression⁶⁹ and elevated serum troponin levels.²³ In addition to documentation of left atrial and ventricular distention, echocardiography may demonstrate abnormal left heart function. The combination of left heart dysfunction and diversion of aortic flow into the lungs (a “ductal steal”) compromises systemic cardiac output, with consequent risk of end-organ ischemia. Increasing ductal left-to-right shunting may be associated with Doppler ultrasound evidence of decreased, absent, or reversal of diastolic flow in the middle cerebral, superior mesenteric, and renal arteries, as well as the descending aorta.²⁹ Impaired cerebral blood flow has been associated with PDA in several studies, implicating but not proving a causal role for PDA in development of IVH and, possibly, PVL.⁵² Similarly, bowel or kidney ischemia resulting from compromised mesenteric or renal arterial flow may be associated with NEC²⁰ or impaired renal function.⁶⁶ Reversal of aortic diastolic flow in term infants with surgical congenital heart disease has been associated with NEC.⁹ However, no empiric data have demonstrated a direct correlation between estimates of severity of the ductal steal and the risk of ischemic complications in preterm infants with PDA.

Clinical Presentation

The clinical presentation of a preterm infant with a persistent PDA typically begins with recognition of the characteristic coarse systolic murmur, heard best along the left sternal border. The continuous “machinery murmur” typical of PDA in an older child is rarely present. Infants with a very large PDA and substantial pulmonary overcirculation may have no audible murmur. An increase in murmur intensity may reflect increasing flow velocity through a narrowing ductus arteriosus rather than increasing shunt volume. Hemodynamic effects of the large left-to-right shunt may be evident in an increased precordial impulse (owing to the increased left ventricular stroke volume) and in arterial pulses that are prominent, bounding, or palpable where they are not normally, such as in the palms (resulting from diastolic runoff into the low pressure pulmonary circulation). In preterm infants greater than 1000 g, systemic arterial diastolic blood pressures are reduced and pulse pressures may be increased. Among those babies less than 1000 g, reduction in both systolic and diastolic pressures without a widened pulse pressure is more typical. These findings are nonspecific, insensitive, and correlate poorly with echocardiographic findings.⁵⁹ Similar physical findings may be present in infants with an aorto-pulmonary window, large arteriovenous malformation, hemitruncus, and certain other cardiac defects, or they may simply reflect a relatively hyperdynamic state of another cause. Often, a hemodynamically significant PDA is suspected only because of signs of excessive pulmonary perfusion, such as an increasing Paco₂ and/or alveolar-arterial oxygen gradient, decreasing lung compliance, or inability to wean the infant from supplemental oxygen, distending airway pressure, or positive pressure ventilation.

Diagnostic Evaluation

Patent ductus arteriosus is most readily confirmed by color Doppler echocardiography, which documents anatomic patency and the direction and velocity of ductal blood flow throughout the cardiac cycle (Figure 83-1). Echocardiography also allows characterization of the hemodynamic consequences of the ductal shunt. Findings that correlate with a large shunt include increased left atrial to aortic root diameter (LA : Ao) ratio, large ductal diameter, left ventricular distention, and reduced or reversed diastolic flow in cerebral, renal, or mesenteric arteries or descending aorta. Measures of myocardial dysfunction, such as the velocity of a mitral regurgitant jet, ratio of early passive to late atrial contractile transmitral filling, and left ventricular isovolumic relaxation time, have been proposed as potential indicators of hemodynamic significance.⁴⁴ The utility of individual markers is low, but combination in a composite score appears to increase diagnostic utility. This approach has not been thoroughly validated, but preliminary results suggest that high composite scores may be predictive for development of chronic lung disease.⁵⁸ These results need to be confirmed and extended to other adverse outcomes potentially attributable to PDA, but represent an essential first step toward establishing the role of such measures in identifying candidates for enrollment in management trials and ultimately for targeted interventions.

Serum biomarkers are potential tools for evaluation of hemodynamic significance of PDA, guiding treatment of PDA, and predicting adverse outcomes. Plasma levels of B-type natriuretic peptide (BNP) or NT-pro-BNP, an inactive byproduct of BNP production, are elevated in infants with significant PDA,11,25 correlate with echocardiographic markers of ductal shunting,^22,25,55 and decrease after the ductus arteriosus closes.^22,55 Use of a decrease in BNP levels as a criterion for stopping indomethacin treatment resulted in administration of fewer doses,³ but no difference in other outcomes. NT-pro-BNP levels at 48 hours were higher in infants less than 32 weeks’ gestation with PDA who died or had grades III/IV IVH than in babies with PDA who survived without IVH,²¹ and survivors with higher NT-pro-BNP levels at 48 hours of age had worse neurodevelopmental outcomes.²⁴ Elevated levels of troponin T at 48 hours of age are associated with presence of PDA, correlate with echocardiographic indicators of ductal shunting,²³ predict death or grades III/IV IVH,²¹ and may help identify infants at greatest risk for poor neurodevelopmental outcome at age 2 years.²⁴ There have been no studies evaluating these markers as criteria for intervention to close the ductus arteriosus.

Chest radiographs may demonstrate an enlarged cardiac silhouette, prominent pulmonary vascular markings, or evidence of pulmonary edema. Although these signs may suggest the diagnosis, they are nonspecific and insensitive.¹⁹