Patients and methods

Results

During the period of study, 90 neonates had been diagnosed with persistent pulmonary hypertension of the newborn. We excluded 5 of these patients, in whom the initial echocardiogram had been performed at another institution (Fig. 1). The clinical and haemodynamic characteristics of the 85 remaining patients are shown in Table 1. In 80 patients (94%), there was extrapulmonary right-to-left shunting, either through an oval foramen and/or a persistently patent arterial duct. The atrial shunt was right-to-left, or bidirectional, in 62 (73%) patients. When bidirectional, the atrial shunt was predominantly left-to-right in systole, and right-to-left in diastole. The majority of the newborns with a patent duct had no aliasing of the colour-Doppler velocities, consistent with an unrestrictive shunting. Bidirectional, or exclusively right-to-left ductal shunting, was found in 62 (73%) cases. The pattern of bidirectional ductal shunting was right-to-left in systole, with a left-to-right flow in diastole. Qualitatively, left and right ventricular systolic function was normal, or mildly depressed, in the majority of the patients. In 64 patients (75%), we found images adequate for measurements of left ventricular ejection fraction and stroke volume. Left ventricular ejection fraction was normal or mildly depressed in 87% of the cases, but a decreased index of left ventricular output, inferior to 2 l/min/m², was found in 61% of the patients.

Figure 1. Profile of the study.

Table 1. Characteristics of the 85 patients with persistent pulmonary hypertension.

Using multivariate analysis, exclusive right-to-left ductal shunting was significantly associated with death, having an odds ratio of 7.8, and 95% confidence intervals from 1.2 to 52.8, with a p value equal to 0.04. Right-to-left ductal shunting was present in 13 of 85 patients (15%, Table 1), 3 of whom subsequently died, with alveolar capillary dysplasia being identified on post-mortem examination in 2 of these neonates. In the third patient, there was a right ventricular dependent circulation, and this neonate died with an intracranial haemorrhage while on extracorporeal membrane oxygenation. In the 4 remaining neonates who died, the ductal shunt was bidirectional in 3, while the arterial duct had closed in the remaining case. In 2 patients with a bidirectional ductal shunt, death was due to renal failure and severe perinatal depression, respectively. The last patient, with a bidirectional ductal shunt, died from respiratory failure after 8 days of extracorporeal membrane oxygenation. He was found, at autopsy, to have alveolar capillary dysplasia. The patient with a closed duct had clinical features consistent with alveolar capillary dysplasia, with echocardiographic evidence of systemic pulmonary arterial pressures. He died from respiratory failure while receiving extracorporeal membrane oxygenation. Permission to perform an autopsy was not obtained.

There was a non-significant trend for patients with exclusively left-to-right ductal shunting to have a greater requirement for extracorporeal membrane oxygenation, with odds ratio of 0.13, and 95% confidence intervals from 0.01 to 1.22, giving a p value equal to 0.07. Of the 5 patients with exclusively left-to-right ductal shunting, 4 required extracorporeal membrane oxygenation, even though the index of oxygenation at 15 min was inferior to 40 in all 5 patients, varying from 19 to 39, with a median value of 31. Associated conditions in these four receiving extracorporeal oxygenation at entry into the trial had been meconium aspiration syndrome in 2 cases, persistent pulmonary hypertension in 1 case, and sepsis with respiratory distress in the other patient. Baseline left ventricular function was qualitatively normal in 3 of the 4 patients, and mildly depressed in one. The common clinical feature that led to cannulation to provide extracorporeal membrane oxygenation in the 4 patients was haemodynamic instability, with hypo-tension refractory to increasing volume or inotropic support. All 4 patients survived.

In all, 44 neonates were randomised to inhale nitric oxide at a starting dose of 40–80 ppm. In four patients, baseline data to calculate the index of oxygenation was not available, so they were excluded from further analysis because their response to inhalation of nitric oxide could not be determined (Fig. 1). The clinical and haemodynamic characteristics of the 40 remaining patients are summarised in Table 2. In 28 patients, there was a positive response to inhalation, whereas 12 had failed to respond, having a decrease of less than 20% in the index of oxygenation. Clinical condition, outcome, and echocardiographic data were similar in those who responded or failed to respond to inhalation of nitric oxide (Table 2).

Table 2. Clinical outcome and echocardiographic data of the 40 patients treated with inhaled nitric oxide.

Multivariate analysis of demographic and echocardiographic parameters in the 40 patients assigned to inhale nitric oxide demonstrated that an exclusive left-to-right atrial shunt increased the risk of failing to respond to the therapy, with an odds ratio of 7.46, 95% confidence intervals from 1.23 to 45.1, and a p value equal to 0.028. In the 28 patients who responded positively to inhalation of nitric oxide, the atrial shunt was left-to-right in 6 (21%), whereas in the 12 who failed to respond, the atrial shunt was left-to-right in half (Table 2). All the patients with an exclusively right-to-left atrial shunt responded positively to nitric oxide. The clinical presentation of the 6 who failed to respond in the setting of a left-to-right shunt was meconium aspiration syndrome in 4 cases, pneumonia in 1 case, and persistent pulmonary hypertension in the other patient. This last patient, with an exclusively left-to-right shunt at the atrial level, also had a right-to-left ductal shunt. His left ventricular function was severely depressed, with echocardiographic evidence of a right ventricular dependent circulation. He responded poorly to inhalation of nitric oxide, with persistence of hypoxaemia and haemodynamic deterioration necessitating extracorporeal membrane oxygenation.

Discussion

A comprehensive echocardiographic examination is an integral element of the initial evaluation of newborns with persistent pulmonary hypertension, both in order to exclude structural congenital heart disease, and to assess cardiac function.¹¹ Echocardiography is also a valuable non-invasive method for evaluating the degree of pulmonary hypertension, the extrapulmonary shunt, and ventricular function.^{3, 5, 8–12} In the present study, the majority of the patients had either normal, or only mildly depressed, left and right ventricular systolic function. Several factors can cause biventricular dysfunction in newborns with persistent pulmonary hypertension. These include pulmonary hypertension by itself, an alteration in the left ventricular geometry due to the pressure overloaded right ventricle, hypoxaemia causing generalised myocardial ischeamia, and metabolic acidemia.¹³ As in our study, however, others have found significant depression of left ventricular function in less than one-fifth of patients with persistent pulmonary hyper-tension of the newborn.^{8, 11} Right ventricular dysfunction is probably more common, and is thought to be due to increased afterload,⁸ although rarely estimated in echocardiographic studies because precise assessment is difficult on account of the complex geometry of the morphologically right ventricle. A low left ventricular output has also been found in many patients with persistent pulmonary hypertension of the newborn,^{8, 11} another feature replicated in our cases. Rather than left ventricular dysfunction, this low output may be the consequence of a decreased left ventricular filling due to reduced pulmonary venous return in the setting of high pulmonary vascular resistance and positive pressure ventilation.⁸ Of our patients, 94% had right-to-left extrapulmonary shunting at the level of either the atrial septum or the persistently patent arterial duct, again as reported in other studies.^{4, 6} On rare occasions, hypoxaemia with intrapulmonary shunting and elevated pulmonary arterial pressure may be associated with a left-to-right extrapulmonary shunting, and pulmonary vascular resistance that is substantially subsystemic.⁸

Although echocardiography has become a routine technique for evaluation of newborns with persistent pulmonary hypertension, the predictive value of the various Doppler echocardiographic measurements has not been clearly established. Previous echocardiographic studies of such newborns have usually concentrated on one particular physiologic variable. Several approaches have focused on measurements of pulmonary arterial pressure.^{5, 9, 11} Although suprasystemic pulmonary arterial pressure was found to be associated with death in some patients,⁹ Skinner et al.¹¹ failed to establish any significant correlation between the degree of increased pulmonary arterial pressure and death. Other echocardiographic studies that focused on assessment of the flow of blood to the lungs showed that this correlated with oxygenation.^{10, 12}

In our study, we have shown that, in patients with persistent pulmonary hypertension of the newborn, an exclusively right-to-left ductal shunt is an independent risk factor for death, with an odds ratio of 7.8, and a p value equal to 0.04. The significance of the shunting across the arterial duct, however, remains unknown. First, the diameter of the duct varies widely in newborns,⁸ although most ducts are not restrictive.⁴ Second, shunting across the duct is not the only source of hypoxaemia in neonates with persistent pulmonary hypertension, as there may also be shunting across the atrial septum or within the lungs.⁸ Musewe et al. found that the patients with right-to-left ductal shunting had the highest pulmonary arterial pressures, and persistently right-to-left ductal shunting was often associated with death.⁹ It is reasonable, therefore, to suspect that poor outcome in this subgroup of patients reflects the severity of pulmonary vascular disease, and may include those patients with alveolar capillary dysplasia and pulmonary hypoplasia.

Extracorporeal membrane oxygenation was required in just over a third of our patients, specifically in 31 of 85. Criterions for the initiation of extracorporeal membrane oxygenation included an index of oxygenation superior to 40 with vasodilator therapy or persistent haemodynamic instability with inadequate perfusion despite inotropic support, and was instituted at the discretion of the attending physician.^{1, 7} Patients with exclusively left-to-right flow across the arterial duct had a tendency to require extracorporeal membrane oxygenation more frequently compared to those with a bidirectional or right-to-left ductal flow, with an odds ratio of 0.13, and a p value equal to 0.07. Our five patients with a left-to-right ductal shunt who required extracorporeal membrane oxygenation had hypotension in the setting of normal or mildly depressed left ventricular function as assessed by echocardiography. A possible explanation for this observation is that the source of hypoxaemia in these patients is often shunting within the lungs or across the atrial septum, and their pulmonary vascular resistance may be relatively low.¹⁴ Vasodilatory therapy that aims at lowering pulmonary vascular resistance, therefore, might aggravate redistribution of blood from the systemic to the pulmonary circulation, reduce systemic perfusion, and worsen the exchange of gases within the lungs. Measures to treat pulmonary hypertension in this instance may exacerbate left-to-right shunting, and increase the likelihood of requiring mechanical support.

Inhalation of nitric oxide has been shown in several studies to improve oxygenation and reduce the need for extracorporeal membrane oxygenation in newborns with persistent pulmonary hypertension.^1–7 The results of our present study, however, indicate that an exclusively left-to-right shunt across the atrial septum increases the risk of failing to respond to nitric oxide, with an odds ratio of 7.46, and a p value equal to 0.028. Left-to-right shunting across the atrial septum is usual in newborns with a patent oval foramen and normally compliant ventricles. In persistent pulmonary hypertension of the newborn, a left-to-right atrial shunt associated with a predominantly left-to-right ductal shunt and a normal biventricular function may reflect intrapulmonary shunting. In this subgroup of patients, systemic oxygenation is significantly less improved by inhalation of nitric oxide.¹⁰ Another potential pathophysiologic mechanism that underlies this finding may involve reduced left ventricular compliance, leading to increased left atrial pressure, with a resultant left-to-right shunt across the oval foramen. Decreased left ventricular compliance may occur in persistent pulmonary hypertension of the newborn due to adverse interaction between the ventricles, a leftward shift of the ventricular septum secondary to right ventricular hypertension, decreased left ventricular diastolic filling, and left ventricular systolic dysfunction due to decreased preload, hypoxaemia, and acidosis. Even when left ventricular systolic function is severely depressed in these patients, the right ventricle can maintain systemic output through the patent arterial duct. Selective pulmonary vasodilation with inhalation of nitric oxide in this circumstance may not give the desired clinical response, because the blood flowing across the duct is redistributed away from the systemic circulation towards the lungs, decreasing post-ductal systemic output, and increasing the left atrial pressure.¹⁴

Several technical aspects of echocardiography limit the conclusions of this study. Precise measurements to assess left and right ventricular function are difficult to perform in these small patients due to their unstable clinical condition. The evaluation of left ventricular ejection and output by offline measurements was only feasible in three-quarters of our cases. The measurements of left ventricular volume to calculate the output may not be accurate, since the higher right-sided pressures may distort ventricular shape and septal motion.³ We did not quantify the percentage of right-to-left atrial or ductal shunting in cases of bidirectional shunting. In many of the echocardiographic records, a good quality peak velocity of tricuspid incompetence accurately to permit measurement of systolic pulmonary arterial pressure was not available. The majority of our patients, nonetheless, had a non restrictive and patent arterial duct, as is well recognised in persistent pulmonary hypertension of the newborn.⁴ In these circumstances, the pulmonary arterial systolic pressure must nearly equal the systemic pressure.^{4, 15}

In conclusion, we have described the descriptive and predictive features of echocardiography in persistent pulmonary hypertension of the newborn. A pure right-to-left ductal shunt identified the patients who are at increased risk of death. A pure left-to-right ductal shunt tends to be associated with greater need for extracorporeal membrane oxygenation, and should prompt cautious re-evaluation of the indication for further treatment aimed at increasing pulmonary vasodilation. The direction of flow across the atrial septum helps to distinguish patients who respond poorly to inhalation of nitric oxide, although the significance of this finding needs to be clarified. Detailed Doppler echocardiographic screening of patients with suspected persistent pulmonary hypertension of the newborn may serve to identify groups requiring specific treatment, and establish factors for risk.

Acknowledgement

Alain Fraisse was supported by the Boston Children's Heart Foundation.