Abstract
CONTEXT: Brain injury is the most common long-term complication of congenital heart disease requiring surgery during infancy. It is clear that the youngest patients undergoing cardiac surgery, primarily neonates and young infants, are at the greatest risk for brain injury. Developmental anomalies sustained early in life have lifelong repercussions.
OBJECTIVE: We conducted a systematic review to examine longitudinal studies of cognitive and/or motor outcome after cardiac surgery during early infancy.
METHODS: Electronic searches were performed in Medline, the Cumulative Index to Nursing and Allied Health Literature (Cinahl), and Embase (1998–2008). The search strategy yielded 327 articles, of which 65 were reviewed. Eight cohorts provided prospective data regarding the cognitive and/or motor outcome of infants who had undergone surgery for congenital heart disease before 6 months of age. Two authors, Ms Snookes and Dr Gunn, independently extracted data and presented results according to 3 subgroups for age of follow-up: early development (1 to <3 years); preschool age (3–5 years); and school age (>5 to 17 years). Weighted analysis was undertaken to pool the results of studies when appropriate.
RESULTS: All of the identified studies reported results of the Bayley Scales of Infant Development for children younger than the age of 3. Outcome data as reported by the Bayley Scales were combined for infants assessed at 1 year of age, revealing a weighted mean Mental Development Index of 90.3 (95% confidence interval: 88.9–91.6) and Psychomotor Development Index of 78.1 (95% confidence interval: 76.4–79.7). Additional analysis was limited by a lack of data at preschool and school age.
CONCLUSIONS: With this review we identified a limited number of prospective studies that systematically addressed outcome in patients at the highest risk. These studies consistently revealed cognitive and motor delay in children after cardiac surgery during early infancy. Additional investigation is required to ascertain the consequences of such impairment during later childhood and into adult life.
WHAT'S KNOWN ON THIS SUBJECT:
Survival rates have improved for infants undergoing cardiac surgery in infancy. Unfortunately, neurologic morbidity is a concern for some survivors. Despite growing awareness of morbidity, there have been no systematic reviews that have examined such outcomes in the current surgical era.
WHAT THIS STUDY ADDS:
A systematic review was performed, considering motor and cognitive outcomes after cardiac surgery at up to 6 months of age. Our detailed analysis of the literature provides an important update and a template for future studies.
Over the last 2 decades, advances in surgical technique and intensive care for infants undergoing surgery for congenital heart disease (CHD) have led to a decline in mortality.1 As survival has improved, long-term disability is now considered a key outcome measure for these patients. Brain injury is the most significant complication of surgery for CHD. Neuropathologic specimens2–4 or cerebral imaging techniques such as ultrasound5–7 and MRI8,–,16 have revealed a spectrum of congenital and acquired insults in up to half of the infants who undergo cardiac surgery. Compared with older children, the youngest infants are at greatest risk of brain injury because of the vulnerability of developing white matter to acute changes in perfusion and oxygenation,17 the greater complexity of the surgery performed, and the associated physiology. Although imaging modalities give us valuable insights into the nature and timing of cerebral injury, the paucity of data correlating findings with long-term follow-up studies limits our ability to make predictions of neurologic outcome from these results.
For the past 2 decades, concerns relating to brain injury in children undergoing open-heart surgery, and the paucity of neurodevelopmental follow-up data, have been highlighted.18 Griffin et al19 published a comprehensive review of neurodisability after surgery for CHD between 1980 and 2001. Of the 19 series identified, most concluded that children with CHD were at increased risk for impaired neurodevelopment regardless of their lesion. However, there was inconsistency in findings in the preschool- and school-aged populations, predominantly because of the spectrum of CHD and the age at surgery. The authors reinforced the lack of available longitudinal data regarding cognitive and neurologic status of children with complex CHD.
Several elegant reviews have contributed to our current knowledge regarding neurodevelopmental outcome in the modern era of infant cardiac surgery.20,–,25 However, to our knowledge no systematic analysis of developmental outcome data has been undertaken in these patients, and no analysis has been performed that specifically targets those having surgery during early infancy. The objective of this systematic review was to examine the literature for longitudinal studies of motor and/or cognitive development in infants undergoing cardiac surgery during the first 6 months of life.
METHODS
Eligibility Criteria
Types of Studies
This review included all randomized, controlled trials or prospective observational cohort studies or collective reports of multiple prospective cohorts that met the inclusion criteria for patients and outcome measures.
Patients
Studies of infants undergoing surgical repair or palliation of CHD at up to 6 months of age were considered for inclusion. Studies for which outcomes related to heart transplant or isolated arterial duct ligation were excluded. Reports relating to surgery performed before 1988 were excluded, because earlier surgery would now be considered less representative of the current era with respect to surgical technique, choice of operation, perfusion practice, and perioperative intensive care. Studies for which the authors reported outcomes solely of infants diagnosed with syndromes known to affect motor and/or cognitive outcome, such as trisomy 21 and the 22q deletion spectrum, also were excluded.
Outcome Measures
We included studies for which motor and/or cognitive outcome using a standardized assessment tool were reported. These tools typically produce a developmental quotient or mean assessment score, which is available for comparison to normative data. Nonstandardized methods of evaluating these outcomes were not considered.
Search Strategy
A comprehensive search was undertaken by using Medline, the Cumulative Index to Nursing and Allied Health Literature (CINAHL), and Embase. Search terms included “cardiovascular surgical procedures” or “cardiac surgical procedures” or “heart defects,” “congenital and brain injuries” or “child development” or “developmental disabilities” or “motor skills disorders” or “psychomotor performance” or “psychological tests not syndrome” or “chromosome disorders or chromosome aberrations.” The search was limited to articles published in English from 1988 to 2008. Two authors, Ms Snookes and Dr Gunn, reviewed the initial search yield on the basis of title and abstract, and studies that did not meet the inclusion criteria were excluded. The review authors then examined the full text of the remaining articles to determine if they met the inclusion criteria. Reference lists from the included articles were reviewed for additional studies.
Quantitative Analysis
Cognitive and motor outcome data were pooled into 3 age groups: early development (1 to <3 years); preschool age (3–5 years); and school age (>5 to 17 years). If serial assessments within the specified age group were reported, data from the most recent assessment was used. For example, if a study reported cognitive outcomes at 9 and 18 months, then the outcome at 18 months was used. If serial assessments across our specified age groups were reported, data from the most recent assessment within each age category were used. For each study, mean outcome values and SDs were reported. For studies for which outcomes using the same outcome measure at the same time point were reported, a weighted estimate of mean outcome values was performed. For these studies, heterogeneity was evaluated by using the I2 statistic. The I2 statistic describes the percentage of total variation across studies that results from heterogeneity rather than chance.26 A value of 0% indicates no observed heterogeneity, and larger values show increasing heterogeneity. The possible reasons for heterogeneity were explored by scrutinizing the studies and performing subgroup analyses when appropriate. Authors of individual trials were contacted to obtain clarification of items when necessary.
RESULTS
Description of Studies
The search strategy initially identified 327 references, of which 266 were excluded on the basis of the title and abstract. Six additional articles were identified through hand-searching, of which 2 were excluded because of cross-sectional study design.27,28 Sixty-five publications required more detailed examination, of which 40 were excluded for 1 or more of the following reasons: an absence of developmental assessment tools and/or systematic reporting of outcome measures at or beyond 1 year of age29,–,33; 16 were excluded because of the inclusion of infants whose age at primary surgery was >6 months, from which it was not possible to extrapolate data from the younger subgroups32,34,–,48; 7 were excluded because of an indiscernible surgical period or surgical period before 198849,–,55; and 12 were excluded because of cross-sectional or retrospective study design.37,–,40,42,45,46,56,–,60
The final 25 articles that met all specified inclusion criteria considered only 8 cohorts; as would be expected from longitudinal studies, many of these cohorts generated multiple publications. Pertinent details of the publications that arose from these 8 studies, on which our results are based, are presented in Table 1. Each cohort is discussed in detail below.
Details of Included Studies
Study 1
Study 1 was a single-center randomized, controlled trial of infants who underwent an arterial switch operation for transposition of the great arteries (TGA) with either an intact ventricular septum or a ventricular septal defect. Infants were randomly assigned to receive either low-flow cardiopulmonary bypass or deep hypothermic circulatory arrest, with stratification according to diagnosis and surgeon. Exclusion criteria were a birth weight of <2.5 kg, a recognizable congenital syndrome, an associated extracardiac anomaly of greater than minor severity, previous cardiac surgery, or the need for additional aortic arch reconstruction before the first follow-up. Participants were assessed at 1, 4, and 8 years of age by using the Bayley Scales of Infant Development I (BSID-I), the Wechsler Preschool and Primary Scale of Intelligence-Revised (WPPSI-R) and the Peabody Developmental Motor Scale (PDMS), and the Wechsler Intelligence Scale for Children 3rd Edition, respectively.61,–,72
Study 2
This was a prospective cohort study of neurodevelopmental outcome of infants who underwent the Norwood operation for palliation of hypoplastic left heart syndrome (HLHS). Exclusion criteria included congenital syndromes known to severely affect neurodevelopment, and non–English-speaking families. Participants were assessed at 1 year of age by using the BSID-II.73
Study 3
Study 3 was a single-center prospective cohort study of neurodevelopmental outcome in relation to apolipoprotein E genotype in infants who underwent surgery before 6 months of age. Exclusion criteria included multiple congenital anomalies, a recognizable genetic or phenotypic syndrome other than 22q11 microdeletions, and non–English-speaking families. Subjects were evaluated at 1 year of age by using the BSID-II.74,–,79
Study 4
Study 4 was a single-center randomized, controlled trial of regional cerebral perfusion or deep hypothermic circulatory arrest in neonates with HLHS. Study participants were stratified according to surgeon and the presence of 1 or more factors thought to increase the risk of an adverse outcome, including pulmonary venous obstruction, prematurity (<36 weeks), a birth weight of <2.5 kg, and clinical shock or clinical seizures preoperatively. Exclusion criteria included a birth weight of <2 kg, a gestational age of <32 weeks, an identified genetic syndrome known to be associated with abnormal neurodevelopment, or a history of maternal substance abuse. Development was measured at 1 year of age with the BSID-II.80
Study 5
This was a single-center prospective cohort study of perioperative predictors of neurologic injury in infants undergoing cardiac surgery before 4 months of age. Exclusion criteria were a known chromosomal abnormality, associated extracardiac abnormality, or an abnormal preoperative neurologic examination. All participants were assessed at 1 year of age by using the BSID-II.81
Study 6
This was a single-center prospective cohort study of neurodevelopmental outcome in relation to perioperative predictors in infants who underwent surgery before 6 weeks of age. Infants who weighed <2.5 kg at birth were excluded. All participants were assessed at 8 months of age by using the BSID-I and at 5 years of age by using the WPPSI-R or WPPSI 3rd edition.82,–,85
Study 7
Study 7 was a single-center prospective cohort study of neurodevelopmental outcome in relation to perioperative predictors in infants who underwent an arterial switch operation before 1 month of age. Exclusion criteria included known preoperative brain damage, congenital brain abnormalities, and/or other chromosomal abnormalities. Outcome was evaluated at 3 years of age by using the BSID-II.86
Study 8
Study 8 was a single-center prospective cohort study of the neurodevelopmental outcome of infants who underwent cardiac surgery before 2 years of age. The subgroup of infants who required surgery in the first month of life was reported separately. Exclusion criteria included HLHS, as well as clinical evidence of preexisting neurologic impairment that was attributable to a cause other than the heart defect. Participants were assessed 12 to 18 months after surgery by using the PDMS and the Griffiths Mental Development Scale, and at 5 years of age using the WPPSI-R.30,32,87,88
Cognitive Outcome
Early Development (1 to <3 years)
All 8 studies included cognitive evaluation during early development, generating 18 publications.63–65,71,73,–,84,86,87 After excluding studies with secondary analyses, primary data concerning early cognitive outcomes were reported in 11 remaining publications. Five articles reported cognitive outcome at 1 year of age with the BSID-II Mental Development Index (MDI).73,75,76,80,81 The average MDI weighted by study sample size was 90.3 (95% confidence interval [CI]: 88.9–91.6; I2 = 55.9) (Fig 1). One article reported 1-year cognitive outcome age with the BSID-I MDI,62 3 articles reported cognitive outcome at ∼18 months of age with the BSID-II MDI,82–84 1 article reported cognitive outcome at 2½ years of age with the BSID-II MDI,86 and 1 article reported cognitive outcome at 1 to 2 years with the Griffiths Mental Development Scale.87
Weighted mean BSID-II outcome scores at 1 year of age. Left, PDI; right, MDI. The PDI and MDI have a mean developmental quotient of 100 (SD: 15).
Motor Outcome
Early Development (1 to <3 years)
Motor evaluation during early development was performed in all of the studies, generating 18 publications.63–65,71,73,–,84,86,87 After exclusion of articles reporting secondary analyses, there were 11 publications from 8 studies that reported primary data concerning early motor development. Five articles reported motor outcome at 1 year of age with the BSID-II Psychomotor Development Index (PDI).73,75,76,80,81 The average PDI of these studies weighted according to study sample size was 78.1 (95% CI: 76.4–79.7; I2 = 69.1) (Fig 1). One article reported motor outcome at 1 year of age with the BSID-I PDI,62 3 articles reported motor outcome at ∼18 months of age with the BSID-II PDI,82–84 1 article reported motor outcome at 2½ years with the BSID-II PDI,86 and 1 article reported motor outcome between 1 and 2 years with the PDMS.87
Preschool Age (3–5 years)
Motor evaluation at preschool age using the PDMS was performed in 1 of the 8 studies, generating 2 publications.65,66
School Age (>5 to 17 years)
There were no articles reporting motor outcome at school age.
DISCUSSION
In this comprehensive assessment of available literature, we identified 8 studies that provided objective assessments of motor and/or cognitive outcomes after cardiac surgery during the first 6 months of life. These cohorts considered infants with TGA (3), infants with a functionally univentricular circulation (2), and patients with diverse congenital cardiac malformations (3). All studies assessed motor and cognitive ability at between 1 and 3 years of age, with all but 1 study reporting cognitive outcomes within 1 SD of the test mean. Motor outcomes were typically worse, with 5 articles reporting motor quotients within 1 SD of the expected outcome and 6 reporting scores between 1 and 2 SDs from the expected mean. Although it was possible to combine data from 5 studies by using the same outcome measure at 1 year of age, the results themselves are cautioned by the inclusion of heterogeneous subgroups. The 5 samples showed a mean BSID-II MDI within 1 SD of the test mean and a mean BSID-II PDI ∼2 SDs below the test mean; however, calculations represent 2 cohorts of patients with functionally univentricular circulations and 3 cohorts with a diverse spectrum of congenital cardiac malformations. Accordingly, moderate heterogeneity was calculated with the I2 statistic, which, despite our relatively strict inclusion criteria, is reflective of variations in cardiac diagnoses, as well as outcome variance between samples.
The paucity of longitudinal, long-term follow-up studies limits the interpretation of these findings into the school years. We would have liked to perform an analysis of published results at preschool and school age; however, there were inadequate data for children older than 3 years of age. Cognitive outcomes between 3 and 5 years of age were reported for 3 studies, with only 1 including assessment of gross and fine motor domains. Raw data were not provided in the motor study; however, patients were reported as on the 9th and 4th percentiles for gross and fine motor function, respectively.66 Again, cognitive outcomes were reported within 1 SD of the test mean. Cognitive outcome at school age was reported for only 1 study, with the sample mean being approximately 3 quotient points below the test mean.70 This study included visual-spatial and visual-motor skill testing (including time to complete a grooved pegboard), and although a large proportion of results were of concern, examining these outcomes was not a primary aim of this review. It should be noted that despite concerns of developmental risk spanning decades, there is insufficient longitudinal data of the late motor outcomes for children who undergo surgery in the first 6 months of life. The literature may be considered difficult to interpret given the heterogeneity of CHD diagnoses, and the very variable timing of surgery and follow-up. Indeed, a number of important studies were excluded from our review for these reasons. As such, future research should be performed and evaluated only by first defining the subgroup of interest.
Developmental brain disturbances and injury to cortical and deep nuclear gray matter have been well described after infant cardiac surgery.89–91 Typically, the risk of white matter injury is greatest during the first weeks of postnatal life, during the most active period of synaptogenesis and myelination.17,89,92 The vulnerability of immature oligodendrocytes is further compounded by additional cardiovascular risk factors relating to the underlying lesion and associated physiology and by the fact that the majority of the most complex operations are performed on the sickest infants. Despite these widely recognized facts, a significant percentage of the published literature combines neurodevelopmental outcomes of the youngest patients with older ones. It was for this reason that a number of well-conducted prospective studies could not be included in our systematic review. We chose to review only those patients operated on before 6 months, given that surgery beyond this age tended to largely fall into lower-risk, definitive procedures in infants and children with a more robust and mature brain.
Attention to surgical variables has shaped the developmental outcome literature for children with CHD and has influenced intraoperative practice. The Boston Circulatory Arrest Study led to changes in surgical practice worldwide and stimulated considerable interest in the association between neurodisability and CHD.62 We noted a wide variation among the studies with regard to specific surgical factors such as pH strategy and the use and duration of deep hypothermic circulatory arrest, which have been implicated as potential risk factors for impaired outcome.35,69,87 We recognize that even within the confines of this review, perioperative practice has evolved in a plethora of major as well as subtle ways, and this could influence the findings reported. This is particularly relevant for extrapolation of results in which changes to clinical practice were made during the time period between surgery and outcome assessment for a single cohort. Examples of evolving practice over the past decade include a “U-turn” in some centers back to circulatory arrest, modified surgical techniques (eg, surgical strategies for HLHS), and perinatal and perioperative care in patients at high risk, including the broader use of extracorporeal life support. Patient characteristics themselves may too have evolved, with a tendency toward more complex surgery for patients with lesions previously thought to be inoperable. The paramount message is that impairment-based outcome studies of infants with CHD should be interpreted in light of sample characteristics before, during, and after surgical intervention. Similarly, data describing outcomes of children with a specific diagnosis such as TGA should be extrapolated to other diagnostic groups with educated caution.
It would be convenient to assume that patient characteristics and perioperative care predictably determine functional outcome. However, as the Boston study showed, outcome differences between study arms altered over time, as did the contribution of specific factors such as postoperative seizures.63 It is increasingly appreciated that early environmental factors can have positive and negative effects on the structure of the brain and on early child development.93,94 In other words, the positive predictive validity of assessment tools is limited by the fact that the rate of skill development is not constant and linear, and early delays do not always predict later problems. Significant variability in early development has been observed within normally developing infants over time and among “healthy” infants assessed at the same time point.95 In patients with CHD, this variability is magnified by the burden of repeated hospital admissions and the limitation of assessments occurring only at a single time point. In addition, the administrators of the tools and the tools themselves may influence the outcomes reported. Ferry18 highlighted a higher rate of abnormalities reported by neurologists compared with clinicians directly managing the infants with CHD. Moreover, the sensitivity of identifying patients with developmental delay is poor: a recent study showed that >70% of physicians almost always use clinical assessment without a formal screening tool,96 which may detect less than half of the children in need of developmental services.97 It is important to note that a seemingly normal 1-year-old who had cardiac surgery in early infancy may display significant abnormalities of neurodevelopment at 4 or 8 years of age.64 Also contributing to this variability are limitations of the sensitivity of early developmental scales compared with more established and sophisticated tests of intelligence, behavior, and higher cortical function that are assessed from school age in children who are likely more accustomed to test situations.
The scope of this review was limited by a number of important factors, including the paucity of prospective studies, the fact that most assessments occurred at only 1 time point, the poor sensitivity of early assessment tools, and that the assessment tools themselves changed over time. Thus, the contribution of specific risk factors to outcome could not be examined.
CONCLUSIONS
This systematic review has shown that well-designed longitudinal studies form a rare but potentially valuable basis for research into techniques for reducing the incidence of brain injury or for modifying its long-term impact after cardiac surgery. Within a subgroup of infants receiving cardiac surgery at <6 months of age, cognitive and motor developmental domains were below the expected mean at all ages studied. Our review has shown that at ∼1 year of age, the risk of motor delay was greater than the risk of cognitive disability.
In an evolving field of medical and surgical management, more definitive outcomes are critical for parent counseling and the provision of timely intervention for the individual. Despite widespread awareness and concern regarding neurodisability in survivors of surgery for CHD, this systematic review has identified the absolute need for focused long-term studies to investigate the neurodevelopmental outcome of these infants at high risk.
ACKNOWLEDGMENTS
We thank Drs Catherine Limperopoulos, Annette Majnemer, and Frank Pigula for prompt correspondence regarding data presented in published studies.
Footnotes
- Accepted November 24, 2009.
- Address correspondence to Lara Shekerdemian, MD, MRCP, FRACP, Royal Children's Hospital, Flemington Road, Parkville, Victoria 3052, Australia. E-mail: lara.shekerdemian{at}rch.org.au
FINANCIAL DISCLOSURE: Dr Gunn receives a postgraduate health research scholarship from the Murdoch Children's Research Institute; the other authors have indicated they have no financial relationships relevant to this article to disclose.
- CHD =
- congenital heart disease •
- TGA =
- transposition of the great arteries •
- BSID =
- Bayley Scales of Infant Development •
- WPPSI =
- Wechsler Preschool and Primary Scale of Intelligence •
- PDMS =
- Peabody Developmental Motor Scale •
- HLHS =
- hypoplastic left heart syndrome •
- MDI =
- Mental Developmental Index •
- CI =
- confidence interval •
- PDI =
- Psychomotor Developmental Index
REFERENCES
- Copyright © 2010 by the American Academy of Pediatrics