Infants born with critical CHD rely on a patent ductus arteriosus for survival. In 1973, Coceani et al demonstrated that prostaglandin E1 promoted ductal patency in fetal lambs. Reference Coceani and Olley1 This was followed by several clinical trials in the late 1970s that culminated in Food and Drug Administration approval in 1981. Reference Olley, Coceani and Bodach2,Reference Neutze, Starling, Elliott and Barratt-Boyes3 The original dosage recommendation for prostaglandin E1 was an initial range of 0.05–0.10 µg/kg/minute with a subsequent decrease to a maintenance dose of 0.025 µg/kg/minute. Reference Freed, Heymann, Lewis, Roehl and Kensey4 Since the official approval of prostaglandin E1, a few studies have investigated the efficacy of lower maintenance dose regimens. Reference Akkinapally, Hundalani and Kulkarni5 Despite these investigations, major dosage guidelines continue to recommend higher doses. 6–8
Early studies that evaluated the side effect profile of prostaglandin E1 suggested that respiratory depression was a major adverse effect. Reference Lewis, Freed, Heymann, Roehl and Kensey9 Neurologic side effects, such as seizures, were also noted to be side effects of prostaglandin E1 therapy. Reference Roehl and Townsend10 There is evidence that prostaglandin E1-induced respiratory depression is dose-dependent, Reference Hallidie-Smith11,Reference Kramer, Sommer, Rammos and Krogmann12 although this is not supported by all studies. Reference Singh, Fong, Salmon and Keeton13 Due to concerns for respiratory depression during transport, many infants are electively intubated prior to medical transport upon initiation of prostaglandin E1 infusion. Reference Browning Carmo, Barr, West, Hopper, White and Badawi14 To avoid mechanical ventilation in neonates due to adverse effects, a lower dose of 0.01 µg/kg/minute has been suggested, Reference Singh and Mikrou15 although there has been little published supporting clinical evidence.
In this study, we sought to determine the efficacy and side effect profile of a starting and maintenance dose of 0.01 µg/kg/minute of prostaglandin E1, which is our long-standing institutional practice. We postulated that this low-dose regimen has a high treatment success rate with a low incidence of seizures and of respiratory depression leading to mechanical ventilation.
Patients and methods
Study design
This study was a retrospective chart review of neonates with ductal-dependent critical CHD who were started on a prostaglandin E1 dose of 0.01 µg/kg/minute. Local institutional review board approval was obtained prior to starting this study. All patients were admitted to a single cardiac ICU between April 2009 and February 2019. The starting date for this review coincided with the installation of inpatient electronic medical records, allowing for more comprehensive data collection. We excluded patients who started prostaglandin E1 after day of life seven, received greater than 24 hours of prostaglandin E1 at another facility, had a total dose time of less than 12 hours, were born with a closed ductus arteriosus, and had a diagnosis of critical CHD with heterotaxy. Patients with heterotaxy were excluded due to the frequent combination of pulmonary outflow tract obstruction and pulmonary venous obstruction, making clinical assessment of ductal patency particularly challenging. Patients who were mechanically ventilated for the duration of their prostaglandin E1 treatment and those with severe neurologic disorders that affected spontaneous breathing, such as trisomy 13 and 18, were excluded from the side effect analysis. For those patients who received greater than 28 days of prostaglandin E1 therapy, only the first 28 days of clinical data were recorded and included in the study. The patient cohort was divided into three groups, comprised of obstruction to systemic blood flow, obstruction to pulmonary blood flow, and inadequate mixing (d-transposition of the great arteries).
Baseline clinical characteristics, including sex, gestational age, birth weight, and prenatal diagnosis, were obtained from admission history and physical for each patient. Prematurity was defined as gestational age less than 37 weeks, and low birth weight was defined as birth weight less than 2 kg. Echocardiogram, along with surgical and catheterisation reports, were utilised to determine the final diagnosis and to assign patients into groups. For inadequate mixing patients, echocardiogram reports were reviewed to confirm ductal patency while receiving prostaglandin E1 as the clinical manifestations of d-transposition of the great arteries are impacted by adequacy of intracardiac mixing as well as ductal patency.
In patients who were diagnosed prenatally, prostaglandin E1 was initiated at 0.01 mg/kg/minute soon after birth, typically via umbilical venous catheter. For those diagnosed postnatally, prostaglandin E1 was initiated upon the diagnosis of critical CHD by echocardiogram and route of administration was predominantly via an umbilical venous or other central venous catheter. Peripheral intravenous access was utilised only in those few infants in whom central venous access could not be obtained. The medication administration record was reviewed to determine whether each patient was transported on prostaglandin E1or if it was initiated at the study site. The total duration of prostaglandin E1 therapy and all stoppages or changes in dosage were recorded. Rationale for prostaglandin E1 cessation and survival to discharge were also documented. Patients underwent regular weight-based dose adjustment as part of routine clinical care.
Treatment success was defined as the patient reaching surgical or catheter intervention or receiving 28 days of therapy without the need to increase prostaglandin E1 dose above 0.01 µg/kg/minute. Timing of intervention was based on patient clinical course and characteristics rather than set protocols. Records were reviewed for seizure activity and for terms indicative of respiratory depression, including hypoventilation and apnoea. In our cardiac ICU, respiratory depression was defined as a decline in respiratory effort resulting in a reduction in heart rate below the minimum accepted for age or a drop in pulse oximetry saturation substantially below baseline. After thorough review, these adverse effects were recorded if deemed to be related or possibly related to prostaglandin E1 by the study team. The need to institute or escalate respiratory support and/or initiate therapy with caffeine was noted. The use of caffeine was only noted if it was in response to suspected prostaglandin E1-induced apnoea. The study endpoint was cessation of prostaglandin E1 due to surgery or catheterisation, transport to another facility, death, or 28 total days of treatment.
Statistics
Categorical variables were reported as their frequency (percentage). All normally distributed continuous variables were reported as mean ± standard deviation. All continuous variables with non-parametric distributions were reported as median (interquartile range). Differences among categorical variables were tested via Pearson’s chi-squared or Fischer’s exact test. Differences among means between normally distributed continuous variables were evaluated with Student’s t-test or analysis of variance. Differences among medians between continuous variables with non-parametric distribution were evaluated with Mann–Whitney U-test or Kruskal–Wallis test. Both univariate and multivariate logistic regression were utilised to compare predictor variables with dichotomous categorical outcome variables. Results were considered to be significant at p < 0.05.
Results
Baseline and clinical characteristics
A total of 154 patients met inclusion criteria for the study (Fig 1). Of the 154 patients, 86 (56%) patients were male and 68 (44%) were female. Mean gestational age was 38.2 ± 2.2 weeks. Twenty-four (16%) patients were born before 37 weeks of gestation and eight patients (5%) were born before 34 weeks of gestation. Mean birth weight was 3.01 ± 0.72 kg. Gestational age was unavailable for three patients and birth weight was unavailable for one patient. Prenatal diagnosis of critical CHD was made in 110 (71%) of these patients. Of 44 postnatally diagnosed patients, 23 (52%) were diagnosed on day of life zero, 12 (27%) on day of life one, seven (16%) on day of life two, one (2%) on day of life four, and one (2%) on day of life six. Clinical characteristics by lesion type are presented in Table 1, and primary cardiac diagnosis is presented in Table 2.
Figure 1. Breakdown of inclusion criteria of patients from April 2009 to February 2019 (*one patient excluded from dose increase sub-analysis).
Table 1. Demographics and clinical characteristics by lesion category
PGE1 = prostaglandin E1. Bold indicates p < 0.05
Table 2. Specific cardiac diagnoses
AV = atrioventricular; DORV = double-outlet right ventricle; HLHS = hypoplastic left heart syndrome; IVS = intact ventricular septum; PA = pulmonary atresia; PS = pulmonary stenosis; RV = right ventricle; TGA = transposition of the great arteries; TOF = tetralogy of Fallot; VSD = ventricular septal defect
The 154 patients were treated with 28,845 hours of total prostaglandin E1 therapy. Of the total time, 16,810 hours (58%) were without mechanical ventilation. Median length of therapy was 129 (117) hours with a range of 14–672 hours. While the infusion was uninterrupted in most cases, it was paused and then resumed at a later time in 18 (12%) patients. Prostaglandin E1 infusion was discontinued due to surgical intervention in 109 (71%), catheter intervention in 22 (14%), death in three (2%), and elective withdrawal of care in six (4%). In one case, the patient was transferred to another facility for intervention while still receiving prostaglandin E1. In 13 (8%) cases, the infusion was maintained for greater than 28 days. This included 10 premature patients with low birth weight, one patient with obstructive liver disease that required surgical intervention and stabilisation prior to cardiac intervention, one patient with pulmonary atresia and intact ventricular septum with right ventricle-dependent coronary circulation who was maintained on prostaglandin E1 until cardiac transplant, and one patient who was not considered for surgery due to partial trisomy 15 and partial monosomy 15. This last patient was maintained on prostaglandin E1 while the family sought consultation from other centres. Out of the 144 patients who had their infusions discontinued after an intervention, 124 (86%) patients survived to discharge.
There were a few differences in clinical characteristics between groups. Patients with inadequate mixing had lower rates of prenatal diagnosis (7/16, 44%) when compared with systemic obstruction (66/89, 74%, p = 0.015) and pulmonary obstruction patients (37/49, 76%, p = 0.018). Patients with inadequate mixing also had their infusion discontinued and resumed (7/16, 44%) more often when compared with systemic obstruction (8/89, 9%, p < 0.001) and pulmonary obstruction patients (3/49, 6%, p < 0.001). There were no other statistically significant differences between the three groups.
Outcomes: clinical efficacy
Of the 153 eligible patients, 127 (83%) were started and maintained on a prostaglandin E1 dose of 0.01 µg/kg/minute until the endpoint. Of the 26 patients who had their doses increased, the final dose was less than 0.05 µg/kg/minute in 15, 0.05 µg/kg/minute in five, and greater than 0.05 µg/kg/minute in two patients. Four of the 26 patients were able to return to a rate of 0.01 µg/kg/minute after being transiently escalated to a higher dose.
Fifteen patients had their dose increased due to both echocardiographic findings and clinical factors suggesting ductal constriction. In systemic obstruction patients, these factors included blood pressure gradients, pulse abnormalities between upper and lower extremities, and elevated serum lactate levels. In pulmonary obstruction and inadequate mixing patients, the primary clinical factor driving dose increase was hypoxemia. In six patients, there was an echocardiographic finding of ductal constriction without corresponding clinical signs. In two patients, the dose was increased due to clinical signs of ductal constriction without confirmation by echocardiogram. In three patients, the rationale for dose increase was not delineated in the medical record.
Only one patient developed closure of the ductus arteriosus while on 0.01 µg/kg/minute of prostaglandin E1. This patient had systemic obstruction with a diagnosis of moderate aortic stenosis. The infusion was initially discontinued as it was unclear if this patient had ductal-dependent circulation and this resulted in ductal closure. Shortly afterwards, the patient developed signs of low cardiac output and necrotising enterocolitis and prostaglandin E1 was resumed. The ductus was partially re-opened with a dose of 0.01 µg/kg/minute, but the dose was increased to 0.05 µg/kg/minute to promote further opening. The dose was subsequently weaned back to 0.01 µg/kg/minute, and an echocardiogram revealed a large ductus. However, three days later, repeat echocardiogram revealed closure of the ductus while on a dose of 0.01 µg/kg/minute.
A greater proportion of patients with pulmonary obstruction (15/49, 31%) required a dose increase than those with systemic obstruction (9/88, 10%, p = 0.003). Additionally, patients with dose escalation had significantly longer treatment times with prostaglandin E1 with a median of 205 (284) hours compared with 117 (104) hours for those without dose escalation (p = 0.002). Univariate (p = 0.002, odds ratio 1.003, 95% confidence interval 1.001–1.005) and multivariate (p = 0.006, odds ratio 1.003, 95% confidence interval 1.001–1.005) logistic regression also identified a statistically significant association between dose escalation and each additional hour of prostaglandin E1 treatment. Univariate and multivariate analysis of demographic and clinical factors with respect to dose escalation are presented in Table 3. One patient with systemic obstruction was excluded from the dose increase analysis as their infusion was discontinued but then subsequently re-initiated at a higher dose, without first resuming low-dose prostaglandin E1.
Table 3. Characteristics of patients who required dose increase
CHD = congenital heart disease; CI = confidence interval; OR = odds ratio; PGE1 = prostaglandin E1
*Adjusted for transport on prostaglandin, dosage interruptions, prenatal diagnosis, duration of therapy, and type of CHDBold indicates p < 0.05
Outcomes: adverse effects
Of the 137 patients analysed for respiratory depression, 38 (28%) had documented respiratory depression at a dose of 0.01 µg/kg/minute. In 10 of these patients, the respiratory depression was transient and did not merit initiation of respiratory support although four were started on caffeine. Fourteen patients (10%) were started on nasal cannula or high-flow nasal cannula, three patients (2%) were placed on continuous positive airway pressure, and 11 patients (8%) were mechanically ventilated via endotracheal intubation as a result of respiratory depression. Of these 11 patients, five were intubated prior to transport and six were intubated after transport. Fourteen patients (10%) were started on caffeine therapy as a result of prostaglandin E1-induced respiratory depression, of which nine were premature infants. Of note, five patients without side effects at a dose of 0.01 µg/kg/minute experienced respiratory depression after dose increase of above 0.01 µg/kg/minute. No infants were intubated during transport. Seventeen patients were excluded from the respiratory side effect analysis: 12 patients were mechanically ventilated for the entire duration of their prostaglandin E1 treatment course and the remaining five patients were excluded due to severe neurologic disorders that predisposed them to intrinsic respiratory depression.
Premature infants were more likely to experience respiratory depression (12/18, 67%) than term infants (26/117, 22%, p < 0.001). Mechanical ventilation was also more frequent in premature infants (6/18, 33%) than in term infants (5/117, 4%, p = 0.001). Univariate logistic regression identified a statistically significant negative association between respiratory depression and both weekly increase in gestational age (p < 0.001, odds ratio 0.636, 95% confidence interval 0.500–0.808) and each additional kilogram of birth weight (p = 0.002, odds ratio 0.410, 95% confidence interval 0.234–0.718). There was significant correlation between gestational age and birth weight, and multivariate logistic regression illustrated that gestational age (p = 0.009, odds ratio 0.633, 95% confidence interval 0.449–0.893) was a better fit as a risk factor than weight (p = 0.917, odds ratio 0.953, 95% confidence interval 0.387–2.349). Univariate and multivariate analysis of demographic and clinical factors with respect to respiratory depression are presented in Table 4.
Table 4. Characteristics of patients who experienced respiratory depression
CHD = congenital heart disease; CI = confidence interval; OR = odds ratio; PGEI = prostaglandin E1
*Adjusted for birth weight, gestational age, duration of unventilated prostaglandin, and type of CHD
Bold indicates p < 0.05
Both total and lesion-specific values for dose increase and respiratory depression are presented in Figure 2. Only one patient experienced a seizure while on prostaglandin E1 although head ultrasound shortly afterwards revealed a new intracranial bleed and the seizure was attributed to this new bleed rather than the infusion.
Figure 2. Patients who required dose increase or experienced respiratory depression overall and by lesion type. Y-axis represents number of patients. Left bars represent dose increase and right bars represent respiratory depression. Dashed boxes represent the number of patients who required dose increase. Dotted boxes represent the number of patients with respiratory depression. Solid grey boxes represent the number of patients who did not require dose increase. Solid black boxes represent the number of patients who did not have respiratory depression. Per cent of patients that required dose increase or had respiratory depression presented in respective boxes.
Discussion
This study demonstrates that a dose of 0.01 µg/kg/minute of prostaglandin E1 was an effective initial and maintenance dose to secure ductal patency in the majority of infants with ductal-dependent critical CHD. In addition, the incidence of prostaglandin E1-induced respiratory depression requiring mechanical ventilation was low, and most infants were safely transported on this dose without mechanical ventilation.
Efficacy of low-dose prostaglandin E1
Prostaglandin E1 has been utilised for decades to maintain ductal patency in infants with critical CHD. In a landmark multicentre study published in 1981, Freed et al evaluated the use of prostaglandin E1 in 492 patients at 56 centres as part of a protocol sponsored by the manufacturer, Upjohn Company. The study population included patients with cyanotic CHD and acyanotic left heart obstructive disease. More than half of the patients were on 0.1 µg/kg/minute throughout the infusion, though a smaller proportion of patients were able to be gradually weaned to lower doses. They recommended initiation at 0.05 µg/kg/minute with titration down to 0.025 µg/kg/minute for maintenance dosing. Reference Freed, Heymann, Lewis, Roehl and Kensey4 Shortly afterwards, Hallidie-Smith investigated lower doses in 52 patients and concluded that side effects are dose-related and recommended an initial dose of 0.001–0.01 µg/kg/minute. However, only 16 patients in this study received prostaglandin E1 in this low-dose window. Reference Hallidie-Smith11 Kourula et al suggested utilising lower doses after analysis of histopathology revealed less ductal damage with lower doses. Reference Korula, Calder and Neutze16 Kramer et al studied a variety of doses in 91 patients and reported on efficacy and side effect profile. The authors utilised a variety of starting and maintenance doses. They ultimately recommended an initial dose of 0.015 µg/kg/minute although they noted that patients with systemic obstruction may require higher doses. The majority of these patients received an initial prostaglandin E1 dose > 0.01 µg/kg/minute, and over half of the patients in this cohort had inadequate mixing. Reference Kramer, Sommer, Rammos and Krogmann12 Huang et al described their successful experience with an initial dose of 0.02 µg/kg/minute and a maintenance dose of 0.01 µg/kg/minute, although only 25 patients received this regimen and all patients had pulmonary obstruction. Reference Huang, Lin and Huang17 Additionally, Yucel et al provided evidence that maintenance doses as low as 0.003–0.005 can maintain ductal patency, although in their cohort of 95 patients, all were started at doses of 0.03 µg/kg/minute or greater with the mean starting dose being 0.065 µg/kg/minute. The authors also concluded that lesions with systemic obstruction may require higher doses. Reference Yucel, Cevik and Bulut18
Despite these investigations into the efficacy of lower dose prostaglandin E1, many dosage guidelines continue to recommend high starting doses. 6–8 Our current study is the largest retrospective study in the current era and is unique in describing infants who were uniformly started on low-dose prostaglandin E1, even prior to transport, and utilised the same low-dose infusion for maintenance. In this study, 83% of the patients maintained ductal patency on a starting and maintenance dose of 0.01 µg/kg/minute of prostaglandin E1.
In our study, higher rates of dose escalation were observed in patients with pulmonary obstruction when compared with those with systemic obstruction. This is in contrast with other studies that have suggested that patients with systemic obstruction require higher doses. Reference Kramer, Sommer, Rammos and Krogmann12,Reference Yucel, Cevik and Bulut18 One possible explanation for this discrepancy is that our systemic obstruction cohort had a high level of prenatal diagnosis (74%) compared with other studies, where this patient population was typically older at diagnosis, and thus we were able to start prostaglandin E1 earlier in patients with systemic obstruction. Within our study, higher doses were required for those with systemic obstruction who were diagnosed postnatally (5/23, 22%) than for those diagnosed prenatally (4/65, 6%, p = 0.049). Another possible explanation is that the patent ductus arteriosus in patients with pulmonary obstruction may be more tortuous and narrower than in patients with systemic obstruction. Reference Matsui, McCarthy and Ho19 Thus, the ductus in pulmonary obstruction may appear less satisfactory on echocardiography, resulting in dose increase.
The only other risk factor that correlated with dose escalation was length of treatment. The dose increase group had a longer median treatment time (205 [284] hours) than the group that was not escalated (117 [104] hours, p = 0.002). However, while those patients who had their dose increased had longer median treatment times, 13 (50%) of the 26 patients who had a dose increase were increased in the first 24 hours, and 20 (77%) were increased before 117 hours, the median treatment time for those who did not have a dose increase. Furthermore, the odds ratio for each additional hour of prostaglandin E1 was only 1.003. While many patients with low birth weight (12/17, 71%) or prematurity (13/24, 54%) were noted to have treatment times greater than 2 weeks, few infants with low birth weight (3/17, 18%) or prematurity (4/24, 17%) had their dose increased, and neither univariate nor multivariate logistic regression identified either as an independent risk factor for dose increase. Additionally, patients with pulmonary obstruction, the other risk factor identified for dose increase, did not have a longer median dose time than systemic obstruction patients, as depicted in Table 1. Thus, although there was a statistically significant correlation between dose increase and duration of therapy, we do not believe there was a clinically relevant relationship.
Additionally, pausing and resuming the prostaglandin E1 infusion did not appear to be a risk factor for dose increase as those who had their infusion paused and resumed did not exhibit higher rates of dose increase (5/17, 29%) than those whose infusion remained constant (21/136, 15%, p = 0.171). Neither univariate (p = 0.157, odds ratio 2.282, 95% confidence interval 0.728–7.151) nor multivariate (p = 0.064, odds ratio 3.878, 95% confidence interval 0.924–16.275) logistic regression found pausing the infusion to be a risk factor for dose increase. Notably, nearly half of Group III patients had their infusion paused and resumed. No association between pausing and resuming and dose increase was found if Group III patients were removed from this subanalysis (4/10 versus 20/127, p = 0.074).
Time to initial prostaglandin E1 treatment has been shown to correlate with increased dosage. Reference Kramer, Sommer, Rammos and Krogmann12 Generally, infants who are diagnosed prenatally typically began their infusion shortly after birth, while postnatally diagnosed infants typically experienced a larger treatment delay. In our study, the majority of patients (71%) were diagnosed prenatally and were started immediately on prostaglandin E1, which made the effect of time to treatment difficult to assess. When the cohort is examined as a whole, there was no significant difference in the rate of dose increase between patients who were postnatally diagnosed (6/44, 14%) and prenatally diagnosed (20/109, 18%, p = 0.636). However, when systemic obstruction patients were examined in isolation, there was a statistically significant difference in dose increase in postnatally diagnosed patients (5/23, 22%) compared with prenatally diagnosed (4/65, 6%, p = 0.049), suggesting that postnatal diagnosis in patients with systemic obstruction may be a risk factor for dose increase.
Adverse effects
Our investigation of the adverse effects of low-dose prostaglandin E1 focused primarily on respiratory depression and seizures. Comparison regarding respiratory depression between different studies is challenging due to differing methodologies and definitions. Lewis et al reported a respiratory depression rate of 13% in their 492 patients but did not exclude infants who were mechanically ventilated for the duration of their prostaglandin E1 treatment course. Reference Lewis, Freed, Heymann, Roehl and Kensey9 Other studies report higher ranges, some with lower doses, although not all reported whether patients required mechanical ventilation secondary to respiratory depression. Reference Meckler and Lowe20,Reference Cucerea, Simon, Moldovan, Ungureanu, Marian and Suciu21 In their study of 49 patients, Talosi et al recorded a 42% rate of apnoea with a 14% rate of intubation due to respiratory depression at initial doses between 0.025 and 0.05 µg/kg/minute. Reference Talosi, Katona, Racz, Kertesz, Onozo and Turi22 Other studies that have investigated lower dose prostaglandin E1 have found a lower incidence of respiratory depression. Reference Huang, Lin and Huang17,Reference Yucel, Cevik and Bulut18 There is evidence that prostaglandin E1-induced respiratory depression is dose-dependent, Reference Hallidie-Smith11,Reference Kramer, Sommer, Rammos and Krogmann12 although one study does not support this. Reference Singh, Fong, Salmon and Keeton13 In a 1994 study, Singh et al investigated lower doses (<0.01 μg/kg/minute) in 34 patients. The authors noted that 53% of their 34-patient cohort experienced side effects although only five (15%) experienced respiratory depression, and the most common side effect in 21% of their patients was necrotising enterocolitis, which is multifactorial. Reference Singh, Fong, Salmon and Keeton13
Our cohort demonstrated an overall rate of clinically relevant respiratory depression of 20%, and 8% of patients received mechanical ventilation due to respiratory depression. While respiratory depression was not a rare occurrence in our cohort, the degree of respiratory depression resulted in less mechanical ventilation than in a study with conventional dosing. Reference Talosi, Katona, Racz, Kertesz, Onozo and Turi22 Prematurity was the main risk factor for both respiratory depression and need for mechanical ventilation. Lewis et al observed increased rates of apnoea in infants weighing below 2.0 kg. Reference Lewis, Freed, Heymann, Roehl and Kensey9 While univariate analysis in our study supported this, multivariate analysis suggested weight, when controlled for other predictors, was a non-significant factor. Of 130 patients without severe neurologic disorders transported on prostaglandin E1, 106 (82%) were transported without mechanical ventilation and no patient needed an advanced airway placed while being transported. Only one patient developed a seizure due to an intraventricular haemorrhage. No patient in our study required discontinuation of prostaglandin E1 due to side effects.
Limitations
This study is a retrospective chart review with no randomisation of dosing. Low-dose prostaglandin E1 has been the standard of care in our cardiac centre for two decades and thus a historical control group is not available. We acknowledge that patients with inadequate mixing are not all classically ductal-dependent and may not uniformly benefit from prostaglandin E1. However, we included these patients to allow better comparison with prior studies. All but one of the 16 patients in this group had clear echocardiographic evidence of ductal patency while receiving prostaglandin E1. Our study focused on respiratory side effects, and definitions may vary between studies. Even within our own institution, different providers may have had slightly differing criteria. Though we typically define respiratory depression as the interruption in spontaneous respiration associated with secondary physiologic changes, it is challenging in a retrospective study to confirm that this definition was utilised uniformly by all providers. A minority of patients with prematurity or low birth weight may have received respiratory support or caffeine prophylactically without meeting these specific criteria. Other adverse effects of prostaglandin E1 have been reported. We did not seek to measure cutaneous flushing or jitteriness as these are both subjective and of little clinical consequence. Fever, hypotension, diarrhoea, and necrotising enterocolitis are likely to be multifactorial, especially in patients with left heart obstructive disease. We limited our study to patients who received less than 28 days of prostaglandin E1 and thus did not assess for long-term side effects, such as hyperostosis, in this population. Lastly, while ductal morphology may be a factor in dose escalation, not all cases of dose escalation had an echocardiographic evaluation prior to escalation, making this factor difficult to assess.
Conclusion
Our findings suggest that a starting and maintenance dose of 0.01 µg/kg/minute of prostaglandin E1 is effective and safe in maintaining ductal patency in the majority of patients with critical CHD. Patients with pulmonary obstruction are at higher risk of needing escalation of therapy. Postnatally diagnosed patients with systemic obstruction are also at a higher risk of escalation than prenatally diagnosed infants. Respiratory depression remains a concern even on low-dose prostaglandin E1, although fewer infants require mechanical ventilation when compared with historical data in the literature. Premature infants are at higher risk for significant respiratory depression requiring escalation of respiratory support. In this subset, prophylactic tracheal intubation for transport may be warranted. Most other patients may safely be transported on a low dose of prostaglandin E1 while breathing spontaneously. These findings may warrant a reassessment of standard dosing guidelines.
Financial Support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Conflicts of Interest
None.
