Hypothermia for cardiogenic encephalopathy in neonates with dextro-transposition of the great arteries

Abstract

Open in new tabDownload slide

OBJECTIVES

Neonates with dextro-transposition of the great arteries (d-TGA) may experience rapid haemodynamic deterioration and profound hypoxaemia after birth. We report on d-TGA patients with severe acidosis, encephalopathy and their treatment with systemic hypothermia.

METHODS

This study is a single-centre retrospective cohort analysis of newborns with d-TGA.

RESULTS

Ninety-five patients (gestational age ≥35 weeks) with d-TGA and intended arterial switch operation were included. Ten infants (10.5%) with umbilical arterial blood pH > 7.10 experienced profound acidosis (pH < 7.00) within the first 2 h of life. Six of these patients displayed signs of encephalopathy and received therapeutic hypothermia. Apgar scores at 5 min independently predicted the development of neonatal encephalopathy during postnatal transition (unit Odds Ratio 0.17, 95% confidence interval 0.06–0.49, P = 0.001). Infants treated with hypothermia had a more severe preoperative course and required more often mechanical ventilation (100% vs 35%, P = 0.003), treatment with inhaled nitric oxide (50% vs 2.4%, P = 0.002) and inotropic support (67% vs 3.5%, P < 0.001), as compared to non-acidotic controls. The median age at cardiac surgery was 12 (range 6–14) days in cooled infants and 8 (4–59) days in controls (P = 0.088). Postoperative morbidity and total duration of hospitalization were not increased in infants receiving preoperative hypothermia. Mortality in newborns with severe preoperative acidosis was zero.

CONCLUSIONS

Newborn infants with d-TGA have a substantial risk for profound acidosis during the first hours of life. Systemic hypothermia for encephalopathic patients may delay corrective surgery without compromising perioperative outcomes.

INTRODUCTION

A significant number of patients with dextro-transposition of the great arteries (d-TGA) display profound hypoxaemia and acidosis after cessation of placental oxygen supply due to insufficient atrium-level mixing. Patients with restrictive foramen ovale or intact atrial septum and ductal constriction have increased morbidities and mortalities. They are considered to be in critical condition after birth and require immediate care, including prostaglandin E1 infusion and balloon atrial septostomy (BAS) [1, 2].

Severe postnatal acidosis and hypoxia–ischaemia can cause neonatal encephalopathy [3]. Application of therapeutic hypothermia for 48–72 h to near-term and term newborns with asphyxial hypoxic–ischaemic encephalopathy (HIE) improves neurological outcomes, when started within 6 h after injury. Existing studies on hypothermia have excluded newborns with severe congenital malformations [4, 5]. However, recent studies show that extended cooling is feasible in neonates with congenital heart disease (CHD) after perinatal acidosis and cardiac arrest, although evidence for a benefit on neuro-development is yet lacking [6, 7].

Early delays in motor development, long-term behavioural and neurocognitive deficits are common in d-TGA patients, and neuro-developmental outcomes have not improved in recent years [8–10]. Preoperative brain injuries, like white matter injuries and strokes, may be detected with magnetic resonance imaging examinations [10–12]. Severe preoperative acidosis might contribute to perinatal brain injury and is associated with long-term neurological deficits in newborns with d-TGA [13, 14]. The impact of a cardiogenic form of neonatal encephalopathy, caused secondary to impaired foetal-to-neonatal transition due to the cardiac malformation, on perinatal brain development and potential neuroprotective strategies need to be further investigated [6, 10].

Attention is shifting from patient characteristics and perioperative factors towards preoperative risks and optimizing preoperative management in d-TGA patients [10, 15]. Based on our previous description of preoperative hypothermia in 7 newborns with complex CHD who experienced postnatal acidosis and exhibited signs for neonatal encephalopathy [6], we aimed to investigate the frequency of profound preoperative acidosis (defined by blood pH < 7.00) and cardiogenic neonatal encephalopathy in newborns with d-TGA in a retrospective single-centre cohort study. We further sought to identify predictors for these events and compared perioperative co-morbidities, treatment, and outcomes of encephalopathic newborns who received therapeutic hypothermia, with non-acidotic d-TGA patients.

PATIENTS AND METHODS

Study population

We included all newborns diagnosed with d-TGA and at least 35 completed weeks of gestational age at birth, who were planned for arterial switch operation before discharge and born between 1 January 2013 and 31 December 2017. Patients with obstetric complications or foetal acidosis, defined as umbilical arterial blood pH ≤ 7.10, surgical therapy other than arterial switch or primary palliative care were excluded. This retrospective study received approval by the Institutional Review Board (#EA2/069/17). Written informed consent of study participants (or their guardians) was not obtained since all data were collected as part of routine clinical care and have been anonymized for the purpose of analysis.

Diagnosis of severe acidosis and neonatal encephalopathy

The lowest blood pH between birth and cardiac surgery of each patient was considered to diagnose severe preoperative acidosis, defined as blood pH < 7.00 in blood gas analysis. Blood gas analyses including electrolytes and glucose were performed after birth and shortly after cardiopulmonary stabilization in the delivery room or successful BAS. Additional blood gas analyses were performed when indicated by the attending physician. We applied the same criteria as for moderate-to-severe asphyxial HIE to diagnose neonatal encephalopathy in infants with severe preoperative acidosis: clinical signs such as lethargy, stupor or coma in combination with either muscular hypotonia, abnormal reflexes, an absent or weak sucking reflex or clinical seizures, and an abnormal background activity or seizures in amplitude-integrated electroencephalography [4, 6]. Infants without severe acidosis were examined clinically and did not receive electroencephalography. Patients were assigned into 3 groups: (i) no presence of pH < 7.00 (control group), (ii) presence of pH < 7.00 without evidence for neonatal encephalopathy and (iii) presence of pH < 7.00 with evidence for neonatal encephalopathy (hypothermia group).

Clinical management

All patients received prostaglandin E1 after birth or postnatal diagnosis of d-TGA. BAS was performed when atrial mixing was regarded as insufficient based on oxygen saturation and echocardiographic assessment of atrial septal anatomy, judged by the attending paediatric cardiologist. Foramen ovale was considered as ‘severely restrictive’ in cases with complete absent of interatrial blood flow in Doppler echocardiography. Moderate whole-body hypothermia was initiated in acidotic newborns with encephalopathy within the first 6 h of life. A mattress with circulating thermostatically controlled coolant fluid (Tecotherm TS med 200; Teccom, Halle, Germany) was used to maintain the target rectal temperature (33.0–34.0°C) [4–6]. All patients received subsequent arterial switch operation.

Study measures

All data were retrieved from medical files. The 3 groups were compared for differences in demographic, obstetric and perinatal data, duration of hospitalization and mortality. Further comparisons included preoperative co-morbidities and treatments, such as mechanical ventilation, inhaled nitric oxide and inotropic support, anatomic variants of d-TGA and the perioperative course. Results of preoperative brain magnetic resonance imaging and neurological examinations upon discharge were analysed for patients who received hypothermia.

Statistical analysis

Data were analysed using IBM SPSS Statistics, version 24.0 (IBM Inc., Armonk, NY, USA). Descriptive data for continuous variables are presented as median and range and categorical variables as relative frequencies. Differences between groups were analysed using Mann–Whitney U-test for continuous variables and Fisher’s exact test for dichotomous variables. Multivariable analysis was conducted on demographic and perinatal variables that reached significance with P-value ≤0.1 in univariate analysis, using a stepwise forward selection logistic regression model. Odds ratios are presented with 95% confidence interval. Receiver operating characteristics analysis and Youden index were used to determine optimal cut-off values between groups. A P-value of <0.05 was considered as statistically significant.

RESULTS

Ninety-six patients with d-TGA were identified based on the inclusion criteria. One patient with pre-existing acidosis at birth (umbilical arterial blood pH = 6.96) was excluded. Severe preoperative acidosis was diagnosed in 10 out of 95 patients (10.5%), exclusively during the first 2 h of life (median 27.5 min of life, range 10–111 min). All events were associated with postnatal severe hypoxaemia due to inadequate interatrial mixing and therefore considered ‘cardiogenic’. Six (6.3%) patients additionally exhibited signs of encephalopathy, although successful BAS was completed in the second hour of life at the latest.

Patient characteristics and postnatal transition

No differences in demographic or obstetric variables were observed between patients with severe preoperative acidosis and the control group, neither with nor without neonatal encephalopathy. Apgar scores were significantly lower in patients who developed neonatal encephalopathy compared to the control group at 5 min (6 [5–7] vs 9 [5–10], P < 0.001) and 10 min (7 [5–8] vs 9 [6–10], P < 0.001) of life. Cardiopulmonary resuscitation was needed more frequently in patients who developed encephalopathy. It was applied exclusively during the first 2 h of life in encephalopathic patients, and before the age of 11 h in the control group (Table 1).

Variables predicting neonatal encephalopathy

In multivariable logistic regression analysis, Apgar score at 5 min was the only factor associated independently with the development of encephalopathy (unit Odds Ratio 0.17 [95% confidence interval 0.06–0.49], P = 0.001, Nagelkerke's R-squared 0.522). A cut-off value of 7.5 for Apgar scores at 5 min of life had the best combined sensitivity and specificity to predict cardiogenic encephalopathy in receiver operating characteristics curve analysis, with an area under the curve of 0.96 (95% confidence interval 0.91–1.00, P < 0.001) (Fig. 1).

Anatomic variants of d-TGA

The majority of patients had d-TGA with intact ventricular septum (see Supplementary Material, Table S1). The presence of severely restrictive foramen ovale, diagnosed in 3 acidotic infants, was associated with the development of preoperative encephalopathy (P < 0.001).

Hypothermia for encephalopathy

Six patients received systemic hypothermia in accordance with the in-house hypothermia protocol. One patient had to be rewarmed after 19 h due to arterial hypotension, poorly responding to inotropes, whereas 5 patients completed 72 h of cooling, as previously described [6].

Preoperative clinical course and treatment

Encephalopathic newborns receiving hypothermia required more frequent (100% vs 35%, P = 0.003) and longer (5 [4–8] days vs 1 [1–7], P < 0.001) mechanical ventilation, treatment with inhaled nitric oxide (50% vs 2.4%, P = 0.002), inotropic drugs (67% vs 3.5%, P < 0.001) and red blood cell transfusions (66.7% vs 8.2%, P = 0.002) compared to controls (Table 2).

Perioperative course and outcomes

Encephalopathic newborns receiving systemic hypothermia tended to be older at surgery (12 [6–14] days vs 8 [4–59] days, P = 0.088). Postoperative co-morbidities were not increased in encephalopathic newborns after application of hypothermia. The total duration of hospitalization did not differ between encephalopathic infants receiving therapeutic hypothermia and the control group (26 [14–133] days vs 26 [22–33], P = 0.900). Severe preoperative acidosis and encephalopathy were not related with infant death. Two patients with intramural coronary artery, 1 patient with bacterial and fungal sepsis, and 1 premature patient with a birth weight of 2300 grams and multiorgan failure after postoperative extracorporeal membrane oxygenation therapy died in the control group (Table 3).

Neurological outcomes of encephalopathic infants

None of the encephalopathic infants had signs for intracranial haemorrhages in transfontanellar ultrasound or clinical seizures during hospitalization. Postoperative intracranial haemorrhage was detected in 4 patients of the control group. Five of 6 cooled infants received preoperative brain magnetic resonance imaging examinations between days 4 and 7 of life (median day 7), none of them showing lesions typically associated with poor neuro-development after HIE [6]. Neurological examinations were normal at age-appropriate level in all encephalopathic infants upon discharge.

DISCUSSION

In this retrospective analysis, severe preoperative acidosis (pH < 7.00) was observed in 10.5% of d-TGA patients, and 6.3% of all patients exhibited signs for encephalopathy.

In a single-centre study on 340 newborns with d-TGA and intact ventricular septum, preoperative metabolic acidosis, defined as pH < 7.10, was present in 23%, with a trend towards less acidotic events (16% vs 26%) in patients diagnosed prenatally. The authors did not further evaluate for pH values below 7.00 or signs of encephalopathy [16]. Exceptional high regional rates of foetal echocardiography resulted in 82 of our patients (86%) being diagnosed with d-TGA prenatally. Surprisingly, despite the availability of immediate postnatal interdisciplinary care, including cardiac catheterization-based interventions in prenatally diagnosed patients, 10 (12.2%) infants experienced severe acidosis with pH values <7.00. Signs for encephalopathy were present in 7.3% of prenatally diagnosed patients, a marked increase compared to 0.3% in a general newborn population [3]. No pH values <7.00 were observed in outborn patients. However, their hospital records were only partly available for analysis. No patient experienced pH values <7.00 beyond the first 2 h of life. These results show that even instantaneous availability of interdisciplinary delivery room care cannot entirely prevent d-TGA patients from developing severe acidosis and encephalopathy and speed of reaction is crucial in these patients [2]. Foetal cardiac interventions, such as atrial septoplasty or atrial septal stent placement to create a sufficient interatrial communication in utero, early inhaled nitric oxide or rapid connection to cardiopulmonary bypass after the cessation of placental circulation may have the potential to improve oxygenation and prevent acidosis-related cardiogenic encephalopathy. They could thus be future therapeutic options for d-TGA patients with severely restrictive interatrial septum [17].

Preoperative acidosis is associated with adverse neurological events and outcomes in d-TGA patients. In newborns with d-TGA and other complex CHD, profound metabolic acidosis was associated with background pattern deterioration and increased seizure activity in amplitude-integrated electroencephalography during the first 72 h after starting prostaglandin E1 [18]. In a prospective longitudinal study on 60 d-TGA patients who underwent neonatal arterial switch operation with combined deep hypothermic circulatory arrest and low flow cardiopulmonary bypass, severe preoperative acidosis (defined as umbilical venous pH value <7.20), severe postasphyxia syndrome with organ failure and cerebral seizures or repeated severe cyanosis caused by intracardiac mixing problems were associated with adverse neuro-developmental outcomes at the mean ages of 10.5 and 16.9 years [13, 14].

The Apgar score is an accepted method to report the postnatal status of newborns and is used as criterion for diagnosing perinatal asphyxia, besides acidosis in blood [4]. Low Apgar scores are associated with preoperative brain injury in magnetic resonance imaging, neurological comorbidity before hospital discharge and mortality in CHD patients [19–23]. Our results indicate that Apgar scores as early as 5 min after birth can help identify patients at risk of developing profound preoperative acidosis with subsequent encephalopathy during foetal-to-neonatal transition, even those whose blood pH reveals acidosis considerably beyond the first 5 min of life.

A paediatric cardiologist evaluated the oxygen saturation and degree of interatrial mixing by echocardiography after birth or diagnosis of d-TGA in all our patients, to assess which infant requires BAS to provide adequate oxygen levels. In our study, all 3 patients with severely restrictive interatrial blood flow experienced profound acidosis and encephalopathy. Nine out of 10 patients with severe acidosis improved significantly after BAS, yet encephalopathy appeared in 6 of these 9 infants. Successful BAS can rapidly improve oxygenation in neonates with restrictive interatrial communication. However, definite criteria for the indication of BAS are lacking, leading to a high variation in BAS rates between hospitals [24]. The potential impact of BAS on preoperative neurological injury is controversially discussed, but improved oxygenation after BAS seems to attenuate brain injury [11, 12, 24].

Since acidosis and encephalopathy were diagnosed exclusively within the accepted time frame of 6 h after birth for the induction of hypothermia for asphyxial HIE, encephalopathic patients received hypothermia according to our internal HIE cooling protocol [4–6]. The preoperative clinical course of cooled newborns was characterized by an increase in co-morbidities, which could be treated adequately by intensive care measures. Hypothermia did not compromise the effect of low-dose prostaglandin E1 infusion to maintain ductal patency [6]. Although cooled encephalopathic patients tended to be older than controls at cardiac surgery, the time of hospitalization was equal. A longer time to surgical repair increases the risk for major postoperative co-morbidities. This implies a prolonged exposure to cyanosis and a decrease in cerebral tissue oxygenation, raising the question whether potential benefits of hypothermia on neuro-development in encephalopathic newborns can outweigh disadvantages through a delay to surgery [12, 25, 26].

Growing evidence suggests that preoperative risk factors contribute significantly to brain injury in newborns with critical CHD, conceivably even more than cardiac surgery or the use of cardiopulmonary bypass [15, 27]. Blood supplied to the brain of foetuses with d-TGA is derived largely from more deoxygenated blood returning from the caval veins, leading to chronic decrease of cerebral in utero oxygenation, which may impair brain growth [28]. Postnatal acidosis and hypoxia might thus occur as a ‘second hit’ to a brain that is already vulnerable. Hypothermia has recently been proposed as therapy in patients with mild HIE and perinatal arterial ischaemic stroke [29, 30]. Neonates with d-TGA, who are at increased risk of various causes of encephalopathy, such as acidosis, hypoxia–ischaemia and stroke, might particularly benefit from hypothermia, the only neuroprotective treatment option currently available [3].

Limitations

Our study has notable limitations, such as its retrospective character and the single institutional focus, which explains the small number of affected patients in our cohort. There is no comparison group of encephalopathic patients who did not receive hypothermia. Blood gas analyses were obtained based on clinical evaluation rather than a fixed protocol. Only blood pH values were used to diagnose severe acidosis, without consideration of base excess or lactate. Postnatal oxygen saturations have not been included in the study. The decision to perform BAS was based on the clinical judgement of the attending paediatric cardiologist, as there was no internal guideline for the performance of BAS. Different frequencies of intensive care measures between study groups could influence the results, independently of encephalopathy or hypothermia. We were unable to identify and include outborn patients in critical condition who died prior transfer to our departments, potentially leading to an underestimation of frequencies for acidosis and encephalopathy. The lack of long-term neuro-developmental and functional outcome data impedes any statement regarding a potential neuroprotective effect of hypothermia. Finally, analyses were exploratory in nature. Nonetheless, this study is the first to investigate the frequencies of severe preoperative acidosis (pH < 7.00) and signs for neonatal encephalopathy in newborns with d-TGA and to analyse the impact of systemic therapeutic hypothermia on neonatal and perioperative management and outcomes. Further research is required to investigate strategies to minimize risk factors for preoperative brain injury, to implement new adjuvant therapies for these high-risk patients and to assess whether d-TGA patients benefit in neuro-development from extending the current indication for systemic hypothermia. Prospective multicentre trials should be favoured to compensate for the scarcity of these events.

CONCLUSION

Neonates with d-TGA may experience profound acidosis during the first 2 h after birth, due to poor atrium-level mixing, potentially leading to encephalopathy. Systemic hypothermia applied to encephalopathic newborns was associated with increased preoperative co-morbidities but did not affect subsequent cardiac surgery nor perioperative outcomes. Despite notable limitations, our results indicate that preoperative hypothermia could be a future adjuvant therapy for a substantial proportion of critically sick newborns with d-TGA.

SUPPLEMENTARY MATERIAL

Supplementary material is available at ICVTS online.

Conflict of interest: none declared.

Author contributions

Vinzenz Boos: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Resources; Software; Visualization; Writing—original draft; Writing—review & editing. Christoph Bührer: Conceptualization; Formal analysis; Writing—review & editing. Joachim Photiadis: Conceptualization; Formal analysis; Writing—review & editing. Felix Berger: Conceptualization; Formal analysis; Resources; Writing—review & editing.

Reviewer information

Interactive CardioVascular and Thoracic Surgery thanks Jose G. Fragata, Arun Gopalakrishnan and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

 
• BAS

  Balloon atrial septostomy

 
• CHD

  Congenital heart disease

 
• d-TGA

  Dextro-transposition of the great arteries

 
• HIE

  Hypoxic–ischaemic encephalopathy