Persistent pulmonary hypertension of the newborn is a relatively common condition occurring in 7 per 1000 live births and results in mortality of up to 30%.Reference Lipkin, Davidson, Spivak, Straube, Rhines and Chang 1 , Reference Walsh-Sukys, Tyson and Wright 2 The condition is characterised by hypoxaemic respiratory failure secondary to failure of a normal transition of the pulmonary vasculature from a high-resistance foetal state to a low-resistance extra uterine circuit.Reference Haworth and Reid 3 Right and left ventricular function may be compromised as a result of the consequential increase in right ventricular afterload secondary to an increased pulmonary arterial pressure and reduction in left ventricular preload secondary to lower pulmonary venous return.Reference Sehgal, Athikarisamy and Adamopoulos 4 The resultant myocardial dysfunction and low cardiac output in the setting of elevated pulmonary pressures may contribute to the risk of morbidity and mortality in those infants.Reference Evans, Kluckow and Currie 5 – Reference Musewe, Poppe and Smallhorn 7 Inhaled nitric oxide is the clinical standard of care, but up to 40% of infants have a transient response or fail to demonstrate any improvement in oxygenation. 8 , Reference Goldman, Tasker, Haworth, Sigston and Macrae 9 The relative contribution of impairment in myocardial performance to the variable efficacy of nitric oxide is not well understood.
Milrinone is a selective phosphodiesterase 3 inhibitor with pharmacological effects that include relaxation of vascular smooth muscles, enhanced myocardial contractility (inotropy), and improved myocardial relaxation (lusitropy).Reference Silver, Harris and Canniff 10 , Reference LeJemtel, Scortichini, Levitt and Sonnenblick 11 Small case series examining the use of milrinone in this population demonstrate an improvement in clinical parameters and a reduction in the use of nitric oxide.Reference McNamara, Laique, Muang-In and Whyte 12 , Reference Bassler, Choong, McNamara and Kirpalani 13 In addition, the pharmacokinetic properties of milrinone in the setting of persistent pulmonary hypertension in the newborn have been delineated.Reference McNamara, Shivananda, Sahni, Freeman and Taddio 14 The previous studies, however, did not describe the change in myocardial function parameters associated with the use of milrinone in infants with pulmonary hypertension. The use of tissue Doppler imaging and tissue Doppler-derived strain and strain rate measurements are gaining interest in the neonatal population. These newer markers, which include a more comprehensive assessment of right ventricular function, may be more sensitive to changes in myocardial performance than conventional makers such as shortening and ejection fraction.Reference Poon, Edwards, Joshi, Kotecha and Fraser 15 – Reference Nestaas, Stoylen, Sandvik, Brunvand and Fugelseth 17 Therefore, examining the change in myocardial performance using these newer markers warrants exploration.
In our centre, infants with a clinical diagnosis of persistent pulmonary hypertension of the newborn who fail to respond to nitric oxide within 4 hours of commencement undergo a comprehensive echocardiogram to rule out congenital heart disease, assess myocardial function, and the degree of pulmonary hypertension. These infants are then commenced on milrinone in an attempt to augment nitric oxide action and improve myocardial performance. We aim to report the effect of milrinone use in this setting on right and left ventricular function measured using conventional echocardiography markers, tissue Doppler imaging, and tissue Doppler-derived strain and strain rate. We also examine the change in clinical cardiorespiratory parameters over the subsequent 72 hours following milrinone use.
Materials and methods
A retrospective chart review was conducted in the neonatal intensive care unit of the Rotunda Hospital, Dublin, Ireland, between January, 2013 and June, 2014. Infants born after 34 weeks of gestation with a clinical diagnosis of persistent pulmonary hypertension of the newborn who failed to reduce their oxygen requirements below 40%, despite a minimum of 4 hours of nitric oxide therapy at 20 parts per million, were included in this review. Infants with congenital heart disease and known/suspected chromosomal syndromes were excluded from this study. The Rotunda Hospital Research Ethics Committee approved this study.
In our unit, infants born after or at 34 weeks of gestation with persistent hypoxaemia and requiring a fraction of inspired oxygen >60% to maintain an oxygen saturation >90% (deemed an adequate saturation) are diagnosed with persistent pulmonary hypertension of the newborn. These infants are invasively ventilated and commenced on nitric oxide at 20 parts per million after ruling out the presence of pneumothorax. A trial of surfactant is sometimes used at the discretion of the attending neonatologist. Oxygenation index is not used to guide nitric oxide therapy in our unit. All infants receiving nitric oxide undergo invasive blood pressure monitoring using umbilical arterial catheters or peripheral radial artery catheters. Systolic and diastolic blood pressures are maintained above the third centile for each given gestation using a variety of inotropes including dopamine, dobutamine, adrenaline, and noradrenaline. Infants failing to tolerate an inspired oxygen requirement <40% to achieve an adequate saturation after 4 hours of nitric oxide therapy at 20 parts per million are referred to a neonatologist with cardiovascular expertise for consultation. The consultation consists of a clinical evaluation and a comprehensive structural and functional echocardiogram to rule out congenital heart disease, assess pulmonary arterial pressure, and cardiac function and output. A diagnosis of persistent pulmonary hypertension of the newborn was made if the following are identified in the absence of cyanotic congenital heart disease: bidirectional or right-to-left shunting across the patent ductus arteriosus; tricuspid valve regurgitation with estimated right ventricular pressures ⩾2/3 systemic pressures; flat septum or a septum bowing into the left ventricular cavity in systole. All first echocardiograms were reviewed by a paediatric cardiologist to confirm the normal structure of the heart. Infants with evidence of elevated pulmonary arterial pressures and a lack of response to inhaled nitric oxide as described above were commenced on milrinone in an attempt to augment nitric oxide action on the pulmonary vasculature and improve myocardial performance. All infants had a second echocardiogram over the subsequent 24–72 hours in order to assess the response to milrinone therapy.
Milrinone was commenced at an initial dose of 0.50 μg/kg/minute up to 0.75 μg/kg/minute and was continued depending on clinical response. No loading dose was used in this cohort in order to minimise the risk of hypotension. Oxygen requirement and nitric oxide were weaned in responders according to unit protocol: Infants with a response to nitric oxide, defined as an oxygen requirement of ⩽40% to achieve adequate saturation, for 6 hours are deemed suitable for weaning. Initially, nitric oxide was weaned by 5 parts per million every 4 hours until 5 parts per million was reached. The nitric oxide was then weaned by 1 part per million every 2–4 hours until discontinued. If there was an increase in oxygen requirements (>20%) to maintain adequate saturation or a discrepancy in the pre/post ductal saturations, the weaning regime was held for 4 hours.
Echocardiography parameters
The echocardiography studies were performed using either the Vivid I or the Vivid S6 ultrasound scanner (GE Medical, Milwaukee, Wisconsin, United States of America). A comprehensive scan was carried out on all infants according to published guidelines.Reference Mertens, Seri and Marek 18 , Reference Lopez, Colan and Frommelt 19 All clinical studies were digitally stored off line to facilitate archiving and later offline analysis using a dedicated work station (EchoPac, Version 112). The following echocardiography parameters were obtained at each of the evaluation time points using previously described methodologyReference Lopez, Colan and Frommelt 19 : patent ductus arteriosus diameter measured at the pulmonary end; direction of flow and patent ductus arteriosus shunt gradient; right ventricular systolic pressure estimated from the tricuspid regurgitant jet if present using the modified Bernoulli equation and assuming a right atrial pressure of 5 mmHg; left and right ventricular outputs (method described below); left ventricular shortening fraction; intraventricular septal wall shape, from the parasternal short-axis view at the level of the papillary muscles during systole, categorised as normal curvature – indicating that left ventricular pressure is greater than right ventricular pressure – flat – indicating that left ventricular pressure is equal to right ventricular pressure – and bowing into the left ventricle – indicating that left ventricular pressure is less than right ventricular pressure.
Left ventricular output was measured as follows: The aortic root diameter was measured at the hinges of the aortic valve leaflets from the long-axis parasternal view. The velocity time index of the ascending aorta was obtained from measuring the pulsed-wave Doppler from the apical five-chamber view. The cursor was aligned to become parallel to the direction of flow. No angle correction was used, and an average of three consecutive Doppler wave forms was used to estimate the velocity time index. Left ventricular stroke volume was then calculated as the product of the velocity time index and aortic cross-sectional area using the following formula: =π(aortic radius).Reference Walsh-Sukys, Tyson and Wright 2 Left ventricular output was determined as the product of stroke volume and heart rate and indexed to weight in kilograms. Right ventricular output was measured using the same approach with the diameter of the pulmonary outflow tract and the velocity time index measured from the long-axis parasternal view.
In addition, we obtained the following parameters to assess right ventricular function: (i) Tricuspid annular plane systolic excursion, a measure of movement of the tricuspid annulus from base to apex during systole, which reflects global right ventricular function. Tricuspid annular plane systolic excursion was measured based on M-mode echocardiography through the tricuspid annulus (7). (ii) Fractional area change, a measure of the change in right ventricular cavity area from diastole to systole in the four-chamber view. The right ventricular cavity areas at end-diastole and end-systole are traced at the endocardial borders. Fractional area change (%) was then calculated using the following formula: (end diastolic area – end systolic area)/end diastolic area.Reference James, Corcoran and Jain 20
Tissue Doppler imaging of the left and right ventricles was performed using the apical four-chamber view. We aligned the pulsed-wave cursor with the longitudinal plane of motion at all times. The sector width was narrowed to just beyond the walls to increase the frame rate and improve the temporal resolution. On the tissue Doppler traces, we measured peak systolic (s`), early diastolic (e`), and late diastolic (a`) velocities. If the e` and a` waves were fused, we measured the peak velocity of the single wave, which we deemed as a` wave. The average of three to five beats was used for each measurement. Isovolumic contraction and relaxation times and left ventricular systolic and diastolic times were also measured. We calculated the myocardial performance index from the tissue Doppler traces using the following formula: (isovolumic relaxation time+isovolumic contraction time)/left ventricular systolic time.Reference Harada, Tamura, Toyono and Yasuoka 21
Tissue Doppler-derived strain and systolic strain rate of the left and right ventricles were also obtained from the four-chamber view. Sector width was narrowed to maximally increase the frame rates. Offline analysis was performed to measure longitudinal peak systolic strain and peak systolic strain rate in the basal segments of the left and right ventricular free wall and the intraventricular septum. A single elliptical region of interest was determined with a width of 3 mm and length of 1 mm. Strain length, the computational distance, was set at 10 mm. These settings have been demonstrated to be the most reliable in term infants.Reference Poon, Edwards, Joshi, Kotecha and Fraser 15 – Reference Nestaas, Stoylen, Sandvik, Brunvand and Fugelseth 17 Event timing, including aortic and mitral valve opening and closure, was determined using the electrocardiogram and pulsed-wave Doppler of the flow across those valves.
Clinical data collection
The following data were collected from the healthcare records for a period of 72 hours after treatment initiation: time of milrinone treatment commencement, initial dose, maximum dose, and duration of milrinone treatment; dose and duration of nitric oxide treatment; ventilation mode and mean airway pressure; and details of oxygen requirement and of other inotropes. Cardio-respiratory variables were recorded on an hourly basis on all infants receiving invasive ventilator support by the nursing staff. The following haemodynamic variables were collected at baseline, before milrinone treatment commencement, and at pre-defined time points (1, 6, 12, 24, 48, and 72 hours) after the commencement of milrinone treatment: heart rate, systolic, diastolic, and mean blood pressures, and blood gases. Oxygenation index was calculated post hoc from information obtained from the records using the following formula: (oxygenation index=mean airway pressure×fractional inspired oxygen/partial pressure of arterial oxygen×7.5×100). An inotrope score was also calculated post hoc to determine the level of vasoactive support during each time point using the following formula: [(dopamine dose in µg/kg/minute)+(adrenaline dose in µg/kg/minute×100)+(noradrenaline dose in µg/kg/minute×100)]. The following outcome data were collected: ventilation days, duration of hospital stay, and death before discharge. Dobutamine was not used concomitantly with milrinone during the study period.
Outcomes of interest
The primary outcomes were tissue Doppler-derived velocities, strain, and systolic strain rate of the right and left ventricles. Secondary outcomes included oxygenation index, blood pressure, heart rate, oxygen requirement, nitric oxide dose, urine output, pH, and the use of other inotropes over the subsequent 72 hours of treatment
Statistical analysis
Data were summarised as means (standard deviation), medians [inter-quartile range], or count (%), unless otherwise stated. Paired data were analysed using the paired t-test for normally distributed data or the Wilcoxon Signed Ranks Test for skewed data. Serial data were tested using one-way repeated measures analysis of variance. If significant, data were compared with respect to the baseline data. A total of two group comparisons were carried out using the Mann–Whitney U test. χ2, or the Fisher exact test where appropriate, was used to compare categorical data. We accepted a p–value <0.05 as significant. We used SPSS (version 21, IBM, New York, NY 10022, United States) to conduct the analysis.
Results
Infant clinical characteristics
A total of 17 infants with a diagnosis of persistent pulmonary hypertension of the newborn were identified as non-responders who were commenced on milrinone treatment following an echocardiogram examination. Their mean (standard deviation) [Range] gestational age and birth weights were 39.8 (2.0) [35.7–42.0] weeks and 3.45 (0.39) [2.71–4.05] kg, respectively. Their median 1- and 5-minute Apgar scores were 5 [3–9] and 7 [5–9], respectively. Among all, eight (47%) were male, and 10 (59%) were delivered by caesarean section; six infants (35%) had a diagnosis of moderate or severe hypoxic ischaemic encephalopathy and were undergoing therapeutic hypothermia; five (26%) had meconium aspiration syndrome; three infants (18%) developed pulmonary hypertension of the newborn associated with sepsis, polycystic kidney disease (and pulmonary hypoplasia), and pneumothorax, respectively; and three infants (18%) had idiopathic pulmonary hypertension of the newborn. Four infants received cardio-pulmonary resuscitation at birth, three required adrenaline, and 10 infants received surfactant therapy before the commencement of nitric oxide therapy. All infants received nitric oxide within 3 hours of birth. The median duration of invasive ventilation for the cohort was 5 [5–8] days. The median length of hospital stay in the tertiary neonatal intensive care unit was 12 [11–16] days. An infant born at 355/7 weeks of gestation died before discharge due to polycystic kidney disease, prolonged anhydramnious, and hypoplastic lungs. None of the infants required extracorporeal membrane oxygenation.
Baseline cardiorespiratory and pulmonary arterial pressure characteristics
The cohort underwent the first echocardiogram 15 [8–28] hours after the commencement of nitric oxide treatment. Milrinone treatment was commenced at a median time of 1 [0.5–3] hour after the initial echocardiogram. The median [range] starting dose was 0.50 [0.3–0.66] µg/kg/minute. Most infants remained on the median dose for the duration of the treatment, with the exception of one infant requiring an escalation to 0.66 µg/kg/minute and two infants up to 0.75 µg/kg/minute. The median duration of treatment was 88 [65–118] hours. The first echocardiogram demonstrated low peak systolic pressure gradient across the patent ductus arteriosus and abnormal septal morphology consistent with high pulmonary arterial pressure (Table 1). The administration of milrinone led to a fall in right ventricular systolic pressure, which was clinically relevant but statistically non-significant, and an increase in the pressure gradient across the ductal shunt (p=0.01) (Table 1). There was a change in the morphology of the septal wall, suggestive of lower pulmonary arterial pressures (p=0.008).
Table 1 Cardiorespiratory characteristics and echocardiography markers of pulmonary hypertension before and after milrinone treatment.
LV=left ventricle; PDA=patent ductus arteriosus; RVSP=right ventricular systolic pressure
Data are presented as means (SD), medians (IQR), or count (%). No statistical analysis was carried out on RVSP as not all infants had tricuspid regurgitant jets over the two time points
Effect of milrinone on clinical cardiorespiratory characteristics
The use of milrinone was associated with a reduction in oxygen requirement after 24 hours of drug initiation (p=0.003) and peaking at 72 hours (p<0.001) (Table 1, Fig 1). There was a concomitant reduction in the dose of nitric oxide by 48 hours (p=0.02), an increase in urine output, and an increase in the arterial pH (Fig 1). There was a reduction in the oxygenation index with a peak effect after 24 hours (p=0.008), and a concomitant reduction in partial pressure of arterial oxygen (p=0.02) and partial pressure of arterial carbon dioxide (p=0.01), with a peak effect at 6 hours (Table 2). In addition, the administration of milrinone was associated with a reduction in systolic, diastolic, and mean blood pressures (p=0.04), peaking at 6 hours after administration. This was associated with a significant increase in the use of vasopressor inotropes at 6 and 24 hours (Fig 2); however, blood pressure began to increase after 12 hours of milrinone administration with a peak at 72 hours (systolic p=0.02, mean p=0.03, diastolic p=0.02), in spite of a reduction in the use of vasopressor inotropes over the same time period. Lactate levels demonstrated a non-significant fall after 12 hours. There was no change in heart rate over the study period (Fig 2).
Figure 1 Change in oxygen requirement, iNO, urine output, and pH after milrinone administration. The x-axis represents the time before milrinone (Time 0) and after milrinone administration in hours. The error bars represent standard error of the mean. *represents a p–value <0.05 compared with baseline (repeated measures analysis of variance (ANOVA)).
Figure 2 Change in blood pressure, heart rate, inotrope use, and lactate levels after milrinone administration. The x-axis represents the time before milrinone (Time 0) and after milrinone administration in hours. The error bars represent standard error of the mean. *represents a p–value <0.05 compared with baseline (repeated measures analysis of variance (ANOVA)). No error bars were used for lactate values due to the small number of entries at each time point. Dias=diastolic blood pressure; Mn=mean blood pressure; Sys=systolic blood pressure.
Table 2 Change in OI, PaO2, and PaCO2 in the cohort.
MAP=mean airway pressure; OI=oxygenation index; PaO2=arterial partial pressure of oxygen; PaCO2=arterial partial pressure of carbon dioxide
*p-value<0.05 compared with baseline (repeated measures ANOVA)
Effect of milrinone on myocardial performance
There was an increase in right ventricular basal longitudinal strain (p=0.01), systolic strain rate (p=0.002), and fractional area change (p=0.007) after milrinone administration. These changes in right ventricular function were temporally associated with an increase in right ventricular output (p=0.004). There was no change in right ventricular tissue Doppler velocities or tricuspid annular plane systolic excursion from baseline after milrinone administration (Table 3). Left ventricular tissue Doppler s` and a` waves increased after milrinone administration (p<0.001 and 0.02, respectively). These changes were accompanied by an increase in myocardial performance index (p=0.03) and left ventricular output (p=0.04). Septal longitudinal strain, systolic strain rate, and shortening fraction did not change over the two time points (Table 3).
Table 3 Right and left functional parameters before and after milrinone treatment.
a`=late diastolic Doppler velocity; e`=early diastolic Doppler velocity; s`=systolic tissue Doppler velocity; TAPSE=tricuspid annular plane systolic excursion.
Data are presented as means (SD). LV basal longitudinal strain and strain rate were obtained from the septum
Left and right functional parameters at baseline and during milrinone treatment were compared between infants with hypoxic ischaemic encephalopathy (n=6) and those with other diagnoses (n=11). There was no difference in any of the functional parameters at baseline or during milrinone treatment between the two groups (data not shown).
Discussion
In this retrospective review, we demonstrated that the addition of milrinone to treat infants with refractory persistent pulmonary hypertension of the newborn was associated with lower pulmonary arterial pressure, an improvement in some right and left ventricular functional parameters, and an increase in cardiac output. There was a concomitant improvement in the oxygenation index with reduced need for inspired oxygen and nitric oxide. There was a transient fall in blood pressure, which was maximal at 6 hours after milrinone administration and required increased use of vasoactive inotropes.
Myocardial dysfunction associated with persistent pulmonary hypertension
Right and left ventricular function may be compromised in persistent pulmonary hypertension of the newborn as a result of increased right ventricular afterload, due to the increased pulmonary vascular resistance, and reduced left ventricular preload, due to the reduced pulmonary venous return.Reference Sehgal, Athikarisamy and Adamopoulos 4 The effects of a pressure-loaded, dilated right heart include a shift in the interventricular septum and compression of the left atrium and ventricle, both resulting in decreased filling, and thus cardiac output. The low cardiac output state resulting from reduced left ventricular preload can lead to a fall in blood pressure in infants with persistent pulmonary hypertension of the newborn necessitating the use of vasopressor inotropes such as dopamine and adrenaline. Animal data suggest that these inotropes raise systemic and pulmonary vascular resistance and may further contribute to right ventricular compromise in the setting of persistent pulmonary hypertension of the newborn.Reference Jaillard, Houfflin-Debarge and Riou 22 , Reference Cheung and Barrington 23 Several studies have demonstrated the association of a low cardiac output in the setting of persistent pulmonary hypertension of the newborn with morbidity and mortality.Reference Evans, Kluckow and Currie 5 – Reference Musewe, Poppe and Smallhorn 7
Our group has recently published normative data of right ventricular function parameters in 50 healthy infants on days 1 and 2 of life, including strain derived from speckle tracking, tissue Doppler velocities, tricuspid annular plane systolic excursion, and fractional area change.Reference Jain, Mohamed and EL-Khuffash 24 The cohort in this study had lower fractional area change (21 versus 33%), tricuspid annular plane systolic excursion (7.3 versus 9.2 mm), s` (5.2 versus 6.6 cm/second), e` (5.1 versus 8.0 cm/second), and a` (7.3 versus 8.3 cm/second) values on day 1 when compared with the healthy cohort. Strain values across the two studies were not comparable, as they were derived using different methods (tissue Doppler-derived versus speckle tracking). Following milrinone treatment, there was a recovery of fractional area change and tricuspid annular plane systolic excursion by day 2 to levels comparable with the healthy population; however, tissue Doppler velocity waves remained lower than their healthy counterparts on day 2. We chose the tissue Doppler-derived method for the assessment of deformation indices in this cohort instead of the speckle tracking technique. This was to facilitate the measurement of systolic strain rate, as the tissue Doppler-derived method is more reliable than the speckle tracking method in deriving strain rate values in newborn infants.Reference Poon, Edwards, Joshi, Kotecha and Fraser 15 – Reference Nestaas, Stoylen, Sandvik, Brunvand and Fugelseth 17
Rationale for using milrinone and its effect on myocardial function
Cyclic nucleotide phosphodiesterases are a family of enzymes that hydrolyse the phosphodiester bond in cyclic guanyl monophosphate and cyclic adenosine monophosphate, thereby inhibiting their pulmonary vasodilator properties. Milrinone is a selective phosphodiesterase 3 inhibitor with pharmacological effects including relaxation of vascular smooth muscles, enhanced myocardial contractility (inotropy), and improved myocardial relaxation (lusitropy).Reference Silver, Harris and Canniff 10 , Reference LeJemtel, Scortichini, Levitt and Sonnenblick 11 The use of milrinone is established in neonates and children following cardiac surgery for the prevention of low cardiac syndrome and the treatment of pulmonary hypertension.Reference Hoffman, Wernovsky and Atz 25 , Reference Jain, Sahni, El-Khuffash, Khadawardi, Sehgal and McNamara 26 In animal models and clinical paediatric studies, milrinone demonstrates a synergistic effect when used with nitric oxide in lowering pulmonary vascular resistance.Reference Deb, Bradford and Pearl 27 , Reference Khazin, Kaufman and Zabeeda 28 The use of milrinone in the setting of persistent pulmonary hypertension of the newborn has previously been examined in smaller case series demonstrating a beneficial synergistic effect when used with nitric oxide.Reference McNamara, Laique, Muang-In and Whyte 12 , Reference Bassler, Choong, McNamara and Kirpalani 13
The pharmacological profile of milrinone in persistent pulmonary hypertension of the newborn, its short-term outcomes, and safety profile were recently delineated by our group in an open-label prospective pharmacological study of 11 neonates with persistent pulmonary hypertension carried out at a different centre (The Hospital for Sick Children, Toronto, Canada).Reference McNamara, Shivananda, Sahni, Freeman and Taddio 14 Although the pharmacological study describes a similar cohort of infants, none of the infants overlapped in this study. There was no interval improvement in median oxygen requirement (100 versus 98%, p=0.35) or oxygenation index (37.2 versus 41.6, p=0.53) after nitric oxide administration, before initiation of milrinone. In the pharmacological study, infants received an intravenous loading dose of milrinone (50 μg/kg) over 60 minutes followed by a maintenance infusion (0.33–0.99 μg/kg/minute) for 24–72 hours. The mean (standard deviation) half-life, total body clearance, volume of distribution, and steady state concentration of milrinone were 4.1±1.1 hours, 0.11±0.01 L/kg/hour, 0.56±0.19 L/kg, and 290.9±77.7 ng/ml, respectively. The initiation of milrinone treatment led to an improvement in partial pressure of artrial oxygen (p=0.002) and a sustained reduction in oxygen requirement (p<0.001), oxygenation index (p<0.001), mean airway pressure (p=0.03), and inhaled nitric oxide dose (p<0.001). Although a transient reduction in systolic arterial pressure (p<0.001) was seen following the bolus, there was overall improvement in base deficit (p=0.01) and plasma lactate levels (p=0.04), with a trend towards lower inotrope score. Serial echocardiography revealed lower pulmonary artery pressure, improved right and left ventricular output, and reduced bidirectional or right-to-left shunting (p<0.05) after milrinone treatment. No infants were withdrawn from the study, and there were no cases of intraventricular haemorrhage, electrolyte disturbances, abnormal liver or coagulation profiles, thrombocytopaenia, need for extracorporeal membrane oxygenation, or mortality.
In this study, we demonstrated that milrinone administration was followed by an increase in both right and left ventricular outputs. This was temporally associated with an improvement in right ventricular strain and systolic strain rate, as well as left ventricular tissue Doppler s` and a` velocities in addition to left ventricular myocardial performance index. This differential effect of milrinone on strain, systolic strain rate, and tissue Doppler parameters in the left and right ventricles is interesting and may reflect the physiological changes resulting from milrinone administration. Tissue Doppler velocities are influenced by changes in preload.Reference Amoozgar, Tavakkoli, Ajami, Borzoee and Basiratnia 29 – Reference El-Khuffash, Jain, Dragulescu, McNamara and Mertens 31 Recent animal data suggest that systolic strain is highly influenced by afterload. Strain rate, on the other hand, is less influenced by changes in cardiac loading conditions and is, therefore, a more reliable measure of contractility. The improvement of afterload-dependent parameters in the right ventricle and preload-dependent parameters in the left ventricle may be explained by lower right ventricular afterload, improved ventilation/perfusion matching, and improved pulmonary venous return and left heart preload.
The effect of milrinone on clinical parameters
The administration of milrinone was associated with a reduction in oxygenation index, oxygen requirement, and nitric oxide dose in this cohort. There was a modest but significant fall in the mean blood pressure occurring 6 hours after administration. This was accompanied by an increase in the use of vasopressor inotropes for a period of ~24 hours after this fall. This may have stemmed from the fact that at 6 hours after milrinone administration, many infants approached the 3rd centile of the blood pressure centiles. This fall in blood pressure was associated with an increase in pH, an improvement in urine output, and a fall in lactate levels. It is important to note that, by 72 hours, there was a significant and sustained increase in all the BP parameters when compared with baseline, occurring in spite of a fall in the use of vasopressor inotropes and a continued use of milrinone. No other vasodilatory agent (such as dobutamine) was used with milrinone to avoid excessive hypotension.
Study limitations
This study is limited by its retrospective nature, the small sample size, and the lack of a control group that did not receive milrinone. As such, the echocardiography and clinical changes occurring during the study period cannot be attributed to milrinone alone. We attempted to assess the response to milrinone in infants with hypoxic ischaemic encephalopathy, as these infants may represent a different physiological entity with a variable response to milrinone. We did not detect a difference between these infants and the remainder with other diagnoses; however, due to the small number of patients, we may have missed a real difference. The different diagnoses in this cohort are also a potential limitation. In addition, we were not blinded to the timing of the echocardiograms when performing the offline assessment of the various parameters, and this may have introduced measurement bias.
Conclusion and future direction
The use of milrinone in infants with persistent pulmonary hypertension in newborns with poor/lack of nitric oxide response is associated with improvement in the efficacy of oxygenation and myocardial performance. The inodilator properties of milrinone make it an ideal agent for neonates who are refractory to nitric oxide or have impaired right ventricular systolic performance. A recent Cochrane review illustrated the lack of randomised controlled trials that have investigated the role and benefits/harm of milrinone compared with nitric oxide, and recommend limiting the use of milrinone in persistent pulmonary hypertension of the newborn in newborns to the research setting.Reference Bassler, Kreutzer, McNamara and Kirpalani 32 It is important to systematically investigate the efficacy of milrinone in the setting of persistent pulmonary hypertension of the newborn before widespread dissemination of this treatment.
A more thorough characterisation of myocardial performance in this condition would improve the approach to therapeutic interventions in the clinical and research setting.
Acknowledgements
We would like to acknowledge the exceptional care the NICU nurses provide for those infants.
Financial Support
This project was supported in part by Friends of the Rotunda Research Grant (Reference: FoR/EQUIPMENT/101572) and the Irish Premature Babies Organisation.
Conflicts of Interest
None.
Ethical Standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national guidelines on human experimentation (http://www.dcya.gov.ie/documents/Publications/Ethics_Guidance.pdf) and with the Helsinki Declaration of 1975, as revised in 2008, and has been approved by the institutional committees (The Rotunda Hospital Research Ethics Committee).
