Outcomes in very low birthweight infants with severe congenital heart defect following cardiac surgery within the first year of life Free

Abstract

Open in new tabDownload slide

OBJECTIVES

Very low birthweight (<1500 g, VLBW) infants with severe congenital heart defect (CHD) are at increased risk for perinatal and operative mortality. This study aims to describe morbidity, long-term mortality and neuro-developmental outcome in early childhood in VLBW infants who received cardiac surgery for severe CHD within 1 year after birth.

METHODS

Monocentric observational study on VLBW infants with severe CHD born between 2008 and 2017. Neurodevelopmental impairment at 2 years corrected age was defined as cognitive deficit, cerebral palsy or major neurosensory deficit.

RESULTS

A total of 24 patients were included. Twenty-one (87.5%) infants underwent cardiac surgery with hypothermia during cardiopulmonary bypass (median temperature 30.3°C, interquartile range 27.0–32.0°C) at a median age of 96 (40–188) days. Seven (29.2%, 95% confidence interval 14.9–49.2%) patients died within the first year after cardiac surgery. Survival rates decreased with increasing STAT mortality category of the surgical procedure. Neurodevelopmental impairment at 2 years of corrected age was found in 9 out of 17 (52.9%) surviving infants, with 8 infants (47.1%) presenting with a cognitive deficit or delay and 4 infants (23.5%) being diagnosed with cerebral palsy. Survival without neuro-developmental impairment was 29.2% (n = 7, 95% confidence interval 14.9–49.2%) in the entire study cohort. Eighty percent of the newborns with dextro-transposition of the great arteries, but no patient with univentricular anatomy, survived without neuro-developmental impairment.

CONCLUSIONS

Individual VLBW infants with severe CHD may develop well despite the high combined risk for adverse outcomes. The type of cardiac malformation may affect early- and long-term outcomes.

INTRODUCTION

Congenital malformations and low birthweight are leading causes of infant mortality [1]. Congenital heart defects (CHD) are the most common congenital malformations, and severe lesions affect about 0.1% of newborns worldwide [2]. Very low birthweight (VLBW, <1500 g) infants, who represent 1% of all live births, have an increased incidence of CHD compared to infants of appropriate weight [3, 4].

Mortality rates of VLBW infants with severe CHD ranged from 18% to ∼50% in previous studies, depending on the definition of severe CHD [5–8]. Surviving VLBW infants with severe CHD are at increased risk of comorbidities, particularly bronchopulmonary dysplasia (BPD) [5, 7, 9]. The risk of major neonatal brain injury in VLBW infants with severe CHD is comparable to that of infants without CHD [7, 9].

Surgery for CHD in early infancy is associated with neuro-developmental disorders and delays [10, 11]. In addition, developmental deficits are common in VLBW infants [12]. In low birthweight infants and premature infants in general, the presence of a CHD was associated with poor neuro-developmental outcomes [8, 13–15]. Few studies have investigated neuro-developmental outcomes in VLBW infants who underwent cardiac surgery for complex lesions. Previously studied cohorts either included notable proportions of infants with low-risk lesions such as septal defects or pooled the outcomes of infants of different weight categories [8, 14, 16, 17].

The aim of this monocentric observational study was to investigate morbidity and mortality after congenital heart surgery in VLBW infants with severe CHD. In addition, we sought to evaluate neuro-developmental outcomes at the corrected age of 2 years.

PATIENTS AND METHODS

We included all infants with severe CHD and a birthweight of <1500 g who underwent congenital heart surgery during the first year of life at our institution between 1 January 2008 and 31 December 2017. Severe CHD was defined using the European network of population-based registries for the epidemiological surveillance of congenital anomalies (EUROCAT) criteria [18].

This study received approval by the Institutional Review Board (Charité Berlin, # EA2/069/17, 18 May 2017). Informed consent from the parent/guardian was not obtained, as all data were collected as part of routine clinical care and have been anonymized for this retrospective study.

Data on patient characteristics, neonatal treatment and major morbidities were extracted from medical reports of neonatal units. Treatment details included administration of prostaglandin E1 or inotropes (dopamine, dobutamine, epinephrine, milrinone or norepinephrine), mechanical ventilation, exposure to inhaled nitric oxide and application of surfactant. Major neonatal morbidities were defined as follows: BPD as oxygen requirement, with or without mechanical respiratory support, at postmenstrual age of 36 weeks. Necrotizing enterocolitis (NEC) as any NEC requiring surgical treatment. Retinopathy of prematurity as retinopathy requiring medical (bevacizumab) or surgical (cryotherapy, laser therapy) therapy. Severe brain injury as grade 3 or 4 intraventricular haemorrhage or cystic periventricular leukomalacia. In infants in whom pulmonary artery banding was the first procedure in a staged approach, we considered the subsequent surgical procedure as main cardiac surgery. Surgical procedures were classified in accordance to The Society of Thoracic Surgeons—European Association for Cardiothoracic Surgery (STAT) Congenital Heart Surgery Mortality Categories [19]. Operative data including cardiopulmonary bypass details were recorded. Major postoperative complications were defined as the need for extracorporeal membrane oxygenation, venous thrombosis, NEC and severe brain injury.

Mortality, including cause of death, was analysed using perioperative records, and reports of the physicians involved in the patients’ care. The date of each infants’ last examination reported to our institution was used to proof survival. Surviving infants were eligible for standardized neuro-developmental examinations at the corrected age of 2 years, which included cognitive development using the Bayley Scales of Infant Development (BSID) 2nd or 3rd Edition, depending on the year of the test. For infants who did not participate in comprehensive BSID examinations, data from follow-up visits in our outpatient clinic, and all reports from their general paediatricians on developmental examinations, including standardized German U-Screenings, were analysed. Neurodevelopmental impairment was defined as one or more of the following: cognitive developmental delay, defined as a BSID-II mental developmental index, or BSID-III cognitive composite score of <85, or ‘untestable’, or any cognitive impairment reported by the paediatrician; cerebral palsy of any degree; bilateral blindness; or hearing impairment that could not be corrected by amplification. Favourable outcome was defined as survival without neuro-developmental impairment.

All missing patient information was reported. Statistical analyses were conducted using R studio v1.4.1106 (RStudio, Boston, MA, USA) with R v4.0.4 (The R Foundation for Statistical Computing, Vienna, Austria). Descriptive data were reported as frequency (proportion) for categorical variables and median (interquartile range) for continuous variables. Subgroups for different outcomes were compared using Wilcoxon rank sum test for continuous variables and Fisher’s exact test for dichotomous variables in univariate analysis. A P-value of <0.05 was considered statistically significant.

RESULTS

Twenty-four VLBW infants with a median birthweight of 1288 (1088–1456) g and a gestational age of 30.14 (28.39–32.25) weeks were included. The most frequent types of CHD were dextro-transposition of the great arteries (D-TGA) in 5 (20.8%), tetralogy of Fallot (ToF) in 5 (20.8%) and aortic arch malformations in 4 (16.7%) infants. Four (16.7%) patients had univentricular anatomy. Ten (41.7%) infants had additional cardiac features, mainly ventricular septal defects or valve anomalies (Table 1).

Transition to extrauterine life was uncomplicated in most infants, as confirmed by overall good Apgar scores. Two infants were diagnosed with microdeletions, and 1 infant each had foetal alcohol syndrome and valproate embryopathy. Fifteen (62.5%) infants received prostaglandin E1 in the postnatal period. Nine infants (40.9%) were diagnosed with BPD, 7 of whom met criteria for BPD before cardiac surgery. Other neonatal morbidities were rare (Table 2). Three out of 5 (60.0%) infants with D-TGA received balloon atrial septostomy between 2 and 22 days of life. Three infants (12.5%) received pulmonary artery banding and 4 infants (16.7%) ductus stenting as palliative procedure before the main cardiac surgery.

All patients received cardiac surgery at a median age of 96 (40–188) days and weight at surgery was 3020 (1690–4782) g. Two (8.3%) infants had surgical procedures of STAT mortality category 1, 6 (25.0%) of category 2, 3 (12.5%) of category 3, 11 (45.8%) of category 4 and 2 (8.3%) of category 5 (Table 1). Three (12.5%) infants received normothermic bypass, the remaining 21 (87.5%) infants underwent surgery during hypothermia at 30.3 (27.0–32.0)°C. Postoperative venous thrombosis occurred in 6 (26.1%) infants, including shunt thrombosis and central venous catheter thrombosis, with occlusions ranging from mild to subtotal. Three infants with thrombosis developed chylothorax requiring dietary measures (Table 3).

Overall mortality was 29.2% (95% confidence interval 14.9–49.2%). Four (16.7%) infants died during postoperative hospitalization. Reasons for death were prolonged hypoxia, heart failure, bacterial sepsis and fungal sepsis, in 1 patient each. Three (12.5%) infants died after hospital discharge, 1 each due to cytomegalovirus infection, shunt thrombosis and veno-occlusive disease. No death occurred beyond the first 9 months after cardiac surgery, and the median follow-up time of 17 (70.8%) surviving infants was 56.8 (36.0–80.2) months (Fig. 1).

Neurodevelopmental data were available for 16 (94.1%) out of 17 infants. Eight children (47.1%) had a standardized neuro-developmental assessment at a corrected age of 25 (25–26) months, and the remaining 8 infants had developmental checkups by their general paediatrician. Cognitive results were as follows: 2 infants had normal BSID scores ≥85 and 6 patients had BSID-II or -III scores ≤70 or were untestable in neuro-developmental assessments. No cognitive impairment was diagnosed in 6 out of 8 infants examined by their paediatrician. Cerebral palsy was diagnosed in 4 (23.5%) infants. No infant showed major neurosensory deficits. In summary, neuro-developmental impairment was diagnosed in 52.9% of survivors (Table 4). Overall, 7 infants (29.2%, 95% confidence interval 14.9–49.2%) of the entire cohort had a favourable outcome, and 9 infants (37.5%, 95% confidence interval 21.2–57.3%) survived with neuro-developmental impairment.

In the univariable analysis, favourable outcome and mortality were not associated with prenatal CHD diagnosis, gestational age, birthweight, age and weight at surgery and cardiopulmonary bypass time. Infants were more likely to survive with favourable outcome when diagnosed with D-TGA (80.0% vs 16.7%, P = 0.017). No patient with univentricular anatomy had a favourable outcome. Subgroup analysis showed decreasing survival rates with increasing STAT mortality category (Table 5).

DISCUSSION

Approximately one-third each of VLBW infants who received surgery within the first year of life for severe CHD had either died, presented with neuro-developmental impairment or had a favourable outcome by the age of 2.

Respiratory morbidity, in particular the risk for BPD, is high in VLBW infants with CHD, likely due to prolonged mechanical ventilation and pulmonary oedema [20]. Nine infants (40.9%) in our cohort received oxygen at 36 weeks postmenstrual age, leading to the diagnosis BPD. However, oxygen dependency at 36 weeks was caused by postoperative ventilation in 2 of the infants at that time, rather than exclusively by a chronic lung disease. Previous studies reported BPD rates of 21–37% in VLBW infants and up to 58% in infants weighing <1000 g at birth, in cohorts with varying proportions of infants with simple and severe CHD [5, 7–9, 14]. Presence of CHD increased the risk for BPD in VLBW infants [7]. Rates of NEC, retinopathy and severe brain injury were low in our cohort, and comparable to those previously reported [7, 9, 14]. The differences in the incidence of NEC, retinopathy and severe brain injury between neonates with and without severe CHD appear to be non-existent or small [7].

Mortality was 16.7% during hospitalization and 12.5% after discharge and confined to the first 9 months after cardiac surgery. The Vermont Oxford Network analysed >99 000 VLBW infants in 2006 and 2007. They reported a mortality rate of 44% among 893 infants with serious CHD, with an adjusted odds ratio for mortality of 6.3 for VLBW infants who had serious CHD. Mortality rate was highest in single ventricle defects, including hypoplastic left heart syndrome (85%), compared to 43% for infants with D-TGA, 32% for ToF and 16% for coarctation of the aorta [6]. Among >100 000 VLBW infants born between 1992 and 2012, mortality was 55% in 299 infants with major CHD, compared to 14% in infants without CHD. Mortality rates did not change over time and were ≥70% in complex lesions including hypoplastic left heart syndrome and D-TGA [5]. In a more recent dataset of VLBW singletons born before 32 weeks of gestation between 2007 and 2015, mortality more than doubled in the presence of a severe CHD (18.6% vs 8.9%) [7]. However, these studies report mortality at hospital discharge and do not analyse cardiac surgery (by lesion), or operative survival [5–7]. Increased operative mortality rates have been reported for preterm infants and infants weighting <2500 g at the time of birth or cardiac surgery [21, 22].

Survival probabilities decreased with each increasing STAT mortality category, but due to the small sample size this relationship is not significant. Kalfa et al. [22] reported that early outcomes in infants weighting <2500 g at the time of cardiac surgery were independent of mortality categories. To date, the association between STAT mortality categories and VLBW infant outcomes has not been further investigated.

VLBW infants have low birthweights not solely because of prematurity. Other factors, such as syndromic diagnoses or multiple pregnancies, may contribute to (very) low birthweight in neonates with CHD. Unknown mechanisms might causally relate low birthweight and cardiovascular malformations [23]. The American Heart Association’s scientific statement on neurodevelopment in children with CHD cites prematurity, but not low birthweight, as a risk factor for developmental disorders or disabilities [11]. However, lower birthweight has been shown to be associated with neuro-developmental impairment in CHD patients [10, 11, 15, 24]. Few studies reported neuro-developmental outcomes in VLBW infants who received surgery for severe CHD. Tomotaki et al. presented neuro-developmental outcomes at 3 years of age for 5 VLBW infants who had undergone surgery for severe CHD. Neurological development was borderline to normal in 2 infants with biventricular repair (for ToF and aortopulmonary window, respectively) and poor in 3 infants with univentricular CHD [16]. Pappas et al. reported on neuro-developmental outcomes in infants with birthweight <1000 g and CHD. The majority of 57% of 49 surviving infants had neuro-developmental impairment at 18–22 months of corrected age. However, 47% of these infants had isolated septal defects, and data on cardiac surgery were not provided [14].

The Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network extremely preterm birth outcome model is widely used for prognostication and provides information on cognitive development, gross motor function including cerebral palsy and profound hearing and vision impairment in high-risk premature infants [25]. The same variables were evaluated in the present study, in which 7 of 17 surviving patients had favourable neuro-developmental outcomes at 2 years of age. In particular, children with D-TGA had a surprisingly good outcome. To assign only children who would certainly do well to the favourable outcome group, the threshold for normal BSID test results was set high at 85 points. The BSID-II scores are on average lower than those of the BSID-III [26]. However, since all infants tested with the BSID-II had scores either above 85 or below 70, a reliable classification into favourable or unfavourable outcome is possible. Mild developmental delays may have gone undetected, as half of the surviving children were assessed only by their general paediatricians, who perform less comprehensive examinations. However, this reflects clinical practice, as children with normal developmental outcomes in routine examinations are often not referred for comprehensive testing. Any form of cerebral palsy was considered an unfavourable outcome, as gross motor levels were not available for 3 out of 4 infants. Profound hearing impairment or deafness was not diagnosed in any of the surviving infants, but mild-to-moderate forms of sensorineural hearing loss were not systematically studied. Infants with CHD who undergo cardiac surgery before the age of 1 are at increased risk for sensorineural hearing loss, which could further impede speech acquisition and development [27].

Neurodevelopmental morbidity in infants with CHD is not limited to cognition, cerebral palsy or neurosensory deficits, but may also affect fine motor development, social interaction, attention, behaviour, executive function and core communication skills, including language (receptive and expressive) [8, 10, 11, 15]. Patterns of neuro-developmental delay may change over time in children with severe CHD and may not become apparent until later childhood or adolescence [11]. In addition, congenital syndromes that are more common in children with CHD may negatively affect neurodevelopment [6, 9, 11, 17]. Longitudinal neuro-developmental assessments could help identify delays and optimize outcomes.

Without a normal birthweight population or a cohort of VLBW infants without CHD for direct comparison, the impact of severe CHD on outcomes remains unclear. However, the overall rates of survival and favourable neuro-developmental outcome in this cohort with a mean gestational age of 30 weeks at birth and severe CHD correspond roughly to those expected for premature infants without CHD who receive active treatment after birth at 24 weeks of gestational age [28].

Limitations

This observational study has notable limitations. The small numbers despite a study period of 10 years, the heterogeneity of the patient population and surgical procedures and the lack of standardized follow-up in some patients impede meaningful conclusions. Further aspects of neurodevelopment were not investigated due to the retrospective nature. Neonatal management and surgical techniques have likely changed and improved between 2008 and 2017. Preoperative mortality in infants treated with palliative or curative intent could not be evaluated because various hospitals from surrounding regions and beyond treat VLBW infants after birth before referring them to the heart center. Despite these limitations, this is to our knowledge the first study to report long-term mortality and neurodevelopmental outcome in a cohort of VLBW infants, who received cardiac surgery exclusively for severe lesions. With continued advances in critical and surgical care, more VLBW infants with severe CHD may receive treatment. Beyond survival, comorbidities and neuro-developmental outcomes of these high-risk patients should be studied. Given the low prevalence of different types of severe CHD among VLBW infants, large sample sizes from multicentre trials or networks are needed to accurately assess risks and outcomes and provide guidance for parental counselling and decision-making in these high-risk patients.

Funding

The authors have no source of funding to declare.

Conflict of interest: none declared.

Data Availability Statement

The data underlying this article cannot be shared publicly due to the privacy of individuals that participated in the study. The data will be shared on reasonable request to the corresponding author.

Author contributions

Vinzenz Boos: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Resources; Software; Supervision; Visualization; Writing—original draft; Writing – review & editing. Felix Berger: Conceptualization; Resources; Writing—review & editing. Mi-Young Cho: Conceptualization; Writing—review & editing. Joachim Photiadis: Conceptualization; Writing—review & editing. Christoph Bührer: Conceptualization; Writing—review & editing. Constanze Pfitzer: Conceptualization; Formal analysis; Investigation; Resources; Writing—review & editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Alvise Guariento, Yoshihiro Oshima and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• BPD

  Bronchopulmonary dysplasia

 
• BSID

  Bayley Scales of Infant Development

 
• CHD

  Congenital heart defect

 
• D-TGA

  Dextro-transposition of the great arteries

 
• NEC

  Necrotizing enterocolitis

 
• ToF

  Tetralogy of Fallot

 
• STAT

  The Society of Thoracic Surgeons—European Association for Cardiothoracic Surgery

 
• VLBW

  Very low birthweight