The risk of autism spectrum disorder and intellectual disability but not attention deficit/hyperactivity disorder is increased in individuals with esophageal atresia

Summary

Knowledge of neurodevelopmental disorders such as attention deficit/hyperactivity disorder (ADHD), autism spectrum disorders (ASD) and intellectual disability (ID) in patients with esophageal atresia (EA) is scarce. The aims of this study were to investigate the prevalence and risk of ADHD, ASD and ID in individuals with EA. Data were obtained from four longitudinal population-based registries in Sweden and analyzed using Cox proportional hazards regression. Patients with EA born in Sweden in 1973–2018 were included together with five controls for each individual with the exposure matched on sex, gestational age at birth, birth year and birth county. Individuals with chromosomal aberrations and syndromes were excluded. In total, 735 individuals with EA and 3675 controls were included. Median age at time of the study was 20 years (3–48). ASD was found in 24 (3.9%), ADHD in 34 (5.5%) and ID in 28 (4.6%) individuals with EA. Patients with EA had a 1.66 times higher risk of ASD (95% confidence interval [CI], 1.05–2.64) and a 3.62 times higher risk of ID (95% CI, 2.23–5.89) compared with controls. The risk of ADHD was not significantly increased. ADHD medication had been prescribed to 88.2% of patients with EA and ADHD and to 84.5% of controls with ADHD. Individuals with EA have a higher risk of ASD and ID than individuals without the exposure. These results are important when establishing follow-up programs for children with EA to allow timely detection and consequentially an earlier treatment and support especially before school start.

INTRODUCTION

Esophageal atresia (EA) is a congenital discontinuity of the esophagus affecting 2.43–3.16 neonates per 10 000 live births.¹ After surgical reconstruction of the esophagus, survival rates are nowadays over 90%.² In conjunction with increased survival, the evaluation of long-term outcome has become even more important. A Swedish follow-up program after EA repair at 3 and 6 months, 1, 3–4, 7–8 and 15 years of age was established in 2011. The program has a multidisciplinary approach with a team consisting of a pediatric surgeon, a pediatric pulmonologist and a dietitian. Previous research has predominantly focused on esophageal and respiratory physical dysfunction^3,4 and health-related quality of life (HRQoL).⁵

Infants with EA undergo esophageal repair within the first days of life. In animal experiments, this is a time period of synaptogenesis when the brain is particularly vulnerable to the neurotoxic effect of general anesthetic agents.⁶ Exposure to anesthetics and surgery in experimental animal models during infancy has been shown to cause life-long cognitive dysfunction.^7–9 These findings have raised concerns that exposure to anesthesia and surgery in humans early in life may result in neurodevelopmental impairment. Although some studies have reported an increased risk of neurodevelopmental impairment after exposure to general anesthesia and surgery,^10,11 others do not support these findings.^12,13 A previous study reported an increased risk of brain injury in infants who underwent neonatal surgery for noncardiac congenital anomalies.¹⁴ In a metaanalysis of neurodevelopmental outcomes after neonatal surgery for major noncardiac anomalies, cognitive development delay was found in 23% of the children.¹⁵ It has not been possible to evaluate if the surgery itself may contribute to adverse cogntive development as infants are not exposed to anesthesia without surgery. In a neonatal piglet model, inguinal hernia repair induced significant increases in cell death in the brain compared with anesthesia alone.⁸ Studies specific on neurodevelopmental outcome in children with EA are scarce, based on small sample size and report conflicting results.^15–21

Neurodevelopmental disorders include a broad group of disabilities involving some form of disruption to brain development.²² Attention deficit/hyperactivity disorder (ADHD), autism spectrum disorders (ASD) and intellectual disability (ID) are diagnoses included in this group.²² The etiology is multifactorial including genetic, prenatal and perinatal factors.²² ADHD is characterized by impaired attention and/or motor hyperactivity-impulsivity.²² ASD is defined by impaired social communication/interaction and restrictive repetitive behaviours.²³ ID is defined as during childhood or adolescence appearing reduced intellectual function identified through clinical evaluation or objectively measured as intelligence quotient (IQ) ≤ 2 standard deviations (70 ± 5) and reduced adaptive functioning.²⁴

To our knowledge, studies in ADHD and ASD in patients with EA do not exist and reports on ID in EA patients without chromosomal aberrations and other syndromes are contradictory. Increased understanding in neurodevelopmental disorders such as ASD, ADHD and ID in patients with EA is crucial to support these children and their caregivers and thus to enable good HRQoL.

The aims of this study were to investigate the prevalence and risk of ASD, ADHD and ID in patients with EA.

METHODS

This study was approved by the Swedish Ethical Review Authority (2019–06506; 2021–00777; 2021–04067). Consent was not obtained as the data from the registers were attained and analyzed anonymously.

The study was based on data retrieved from four high-quality nationwide Swedish health data registries. The personal identity number assigned to every Swedish citizen at birth or at immigration was used to link information across registers. The Medical Birth Register (MBR) includes 98% of all births in Sweden. It was established in 1973 and contains maternal data on pregnancies and perinatal data of the newborns. Missing data on deliveries are estimated to be 1–3% in the latest 20 years. The National Patient Register (NPR) was started in 1964 and all hospitals in Sweden have been included since 1987. The register contains information about sex, age, date of admission and discharge. It also contains diagnoses according to the International Classification of Disease (ICD). During the study period, different ICD-versions were used: ICD-8 from 1973 to 1986, ICD-9 from 1987 to 1996 and ICD-10 from 1997 and onwards. The register also keeps codes for surgical procedures. Since 2001, the register covers outpatient visits to a physician, including psychiatric care by both private and public caregivers. The Prescribed Drug Register (PDR) was established in 2005 and contains information on dispensed prescribed medication and drug identity according to the Anatomical Therapeutic Chemical Classification System (ATC). It is regarded to maintain a high quality since the registration process is automated. The Causes of Death Register (CDR) comprises information on death causes since 1961 and, since 1991, includes data on all deceased persons. Missing data on death causes are estimated to be 1–2% in recent years.

Study cohort

All individuals with EA born in Sweden from 1 January 1973 to 31 December 2018 who underwent surgical reconstruction for EA were included. They were identified through the NPR by an ICD code for EA from ICD-10: Q 39.0–Q39.2. ICD-9:750D and ICD-8:750,20 750,28. One of the two inclusion criteria had to be fulfilled to be included in the study cohort.

1. An ICD-code for EA and a surgical procedure code specific for EA.

2. An ICD-code for EA and a hospital stay of at least 7 consecutive days before the age of 30 days.

Through linkage with the NPR, exclusions were made of individuals with chromosomal aberrations diagnosed using the ICD-10: Q90–Q99, ICD-9: 758A–758X, ICD-8: 759,30–759,59 and other syndromes: ICD-10: Q870, Q871, Q873–Q89.

Control cohort

Five controls were drawn at random from the MBR for each individual in the exposed cohort matched on sex, gestational age at birth, birth year and birth county. Individuals with the exposure were excluded through linkage with the NPR. As for the study cohort, exclusion for chromosomal aberrations and syndromes was made.

Outcome variables

From the NPR, data were collected on all in-patient hospital visits up to and including the 5-year birthday. From the age of 5, outcome diagnoses from ICD-9 and ICD-10 and the age at first registered were obtained. Accordingly, individuals born 2016 or later or deceased within 5 years were not included. The following diagnoses were obtained:

Autism spectrum diagnoses (ASD) ICD-10: F84, ICD-9: 299.

Attention deficit hyperactivity disorder (ADHD) ICD-10: F90.0–F90.1, F90.8–F90.9, ICD-9: 314.

ID ICD-10: F70–F79, ICD-9: 317–319.

Information on ATC-codes of prescription of drugs for treatment of ADHD N06B, C02AC02 was obtained from the PDR. Information of deceased individuals was collected from the CDR.

Statistical analysis

Categorical variables are presented as numbers (n) and percentages (%) and continuous variables as median (M) and range. Differences in proportion between cases and controls were analyzed using the Chi-square test. The Mann–Whitney U-test was used to test differences between cases and controls for numerical variables. Data were analyzed using Cox proportional hazards regression where the hazard ratio (HR) indicates the relative difference in risk of outcome between case and control groups. HR was adjusted for sex, year of birth and gestational age. The significance level was set to P < 0.05. For the statistical analyses, R version 4.1.1 (R Foundation for Statistical Computing, Vienna, Austria) was used.

RESULTS

After excluding individuals with chromosomal aberrations and other syndromes, totally 984 individuals with the diagnosis of EA were identified through the NPR. Of these individuals, 619 met inclusion Criteria 1 and 116 fulfilled inclusion Criteria 2, resulting in a study cohort of 735 individuals with EA. The unexposed, control group consisted of 3675 individuals. Characteristics of the study population are presented in Table 1. Median gestational age was 38 weeks and 227 individuals (30.9%) were born preterm (<37 gestational weeks). The proportion of patients who were born small for gestational age (SGA) in the EA cohort was significantly higher, 18% compared with 5% in the control group (P < 0.001). Median age at the time of the study was 20 years (3–48). The prevalence of cardiac malformation was significantly higher, 17.7% in the EA group compared with 0.4% in the control group (P < 0.001). A diagnosis of VACTERL association was found in 55 individuals with EA (7.5%). The mortality rate in the exposed group was 6.1% compared with 1.6% in the control group (P < 0.001) and median age at death was 0 (0–35) and 0 (0–40), respectively.

The proportion of ASD, ADHD and ID in 614 individuals with EA and 3040 controls is shown in Table 2. ASD was found in 24 (3.9%), ADHD in 34 (5.5%) and ID in 28 (4.6%) individuals with EA. The median age at diagnosis was 11.5 years in ASD and ADHD and 10.5 years in ID (Table 2). Age at diagnosis was not significantly different in the EA cohort compared with controls. Among the 28 individuals with ID in the EA group, 18 (64.3%) had a diagnosis of mild ID.

In Table 3, the results of the Cox proportional hazards regression are presented. Compared with controls, individuals with EA had a 1.66 times higher risk of being diagnosed with ASD and a 3.62 times higher risk of ID. The risk of mild ID compared to controls was three times higher. The risk of ADHD was not significantly increased. When dividing the study cohort into two different time periods, individuals born 1973–1996 and 1997–2018, we found that ADHD increased similar to exposed and unexposed individuals. For the later period, the incidence of ADHD was 7.2% in patients with EA and 5.8% in controls. For the period 1973–1996, the incidence in patients with EA was 3.6% and in controls 3.8%.

Patients with EA born at a gestational age < 34 weeks had a 2.45 times higher risk of ID than patients born at gestational age > 36 weeks (95% confidence interval [CI], 1.24–4.85). Patients with EA born between 34 and 36 week of gestation had a 2.04 times higher risk of ID compared with those born at gestational age > 36 weeks (95% CI, 1.19–3.52). No significant relation was found between gestational age and the risk of ASD or ADHD in our study cohort.

ADHD medication had been prescribed for 88.2% of the patients with EA and ADHD and for 84.5% of the controls with ADHD.

DISCUSSION

This is the first population-based cohort study which investigates the risk of ASD, ADHD and ID in individuals with EA. We found that the risk of ASD and ID was increased, but not the risk of ADHD. All three diagnoses are included in a group of neurodevelopmental disorders that have some disruption of the developing brain in common.²²

The most common subtype of EA based on Gross classification is a trachea-esophageal fistula at the lower pouch (85%).²⁵ The duration of the surgical procedure of esophageal repair is around 2 hours. Another type is long-gap EA that may require more complex surgery for reconstruction with longer operating time.²⁶

Animal studies in different species exposed to anesthesia and surgery have provided clear evidence that exposure early in life leads to neuronal cell death in the brain and deficits in cognitive function.^6–9 Single long exposures and multiple exposures adversely affect neurodevelopment. The neurotoxic potential of general anesthesia is dependent of the duration and timing of exposure to general anesthetic agents.²⁷ In an experimental model of piglets, surgery designed to replicate an inguinal hernia repair induced significant increases in cell death compared with anesthesia alone.⁸

In humans, studies in the effect of anesthesia and surgery are conflicting. An explanation might be the inclusion of children with various diagnoses exposed to anesthesia and surgery at various ages, varying durations of surgery, varying study designs and outcome measures, and small sample sizes.^10–13

A recent review reported a prevalence of ASD of 1.7% in the pediatric population.²⁸ In this study, the prevalence of ASD in patients with EA was 3.9% and the risk of ASD was increased compared with the control population. A meta-analysis of perinatal and neonatal risk factors for ASD did not show that general anesthesia was a risk factor for ASD.²⁹ In our cohort of EA patients, there was a significantly higher proportion of individuals born SGA, a known risk factor of ASD.²⁹ Studies in patients who have undergone surgery for cardiac malformations and congenital diaphragmatic hernia report a high prevalence of ASD.^30,31 Fluctuations in blood pressure, arterial CO₂ and duration of anesthesia are perioperative risk factors of brain injury that these diagnoses have in common with EA.³²

The global prevalence of childhood and adolescent ADHD is around 5%,³³ the same prevalence as we found in EA patients and their controls in this study. This is in agreement with a previous study where we found no association between intraabdominal surgery in congenital malformations during infancy and ADHD.³⁴ Also, a previous study of patients with Hirschsprung disease, another congenital malformation, found no increased risk of ADHD.³⁵ In contrast, Ing et al. reported an increased risk of ADHD in children under age 5 with a single exposure to minor surgery requiring anesthesia,³⁶ and, in a recent study,³⁷ the risk of ADHD and ASD in patients with anorectal malformations was increased. Since we used ADHD medication as a proxy for ADHD diagnosis and the proportion of individuals with ADHD medication was lower than those with ADHD diagnosis both in the EA cohort and the controls, it is not likely that we have underestimated the prevalence of ADHD in this study cohort.

In line with previous studies of cognitive function in patients with EA, this study found an increased risk of ID in EA patients compared with healthy individuals.^17,21,38 The risk of ID in our EA patients was even higher than the increased risk of ASD and the risk was related to lower gestational age. In contrast, others have reported normal mental development in children operated on for EA.^16,18–20 However, these studies included only small cohorts ranging from 30 to 59 children with EA. Moreover, the follow-up was from 12 months of age up to 5 years compared with our study with a median age at follow-up of 20 years. The rates of ID may change with longer follow-up, as found in a recent systematic review and meta-analysis of long-term neurodevelopment in children born with EA.³⁹ Infants up to 12 months of age had cognitive function within the normal limits in most studies, whereas two of four studies in school-aged children (6–18 years) found impaired cognitive performance with lower IQ levels.^17,38

The Swedish follow-up program for EA does not include an evaluation by a psychologist. Based on these results, neurodevelopmental screening at minimum using questionnaires to parents and patients should be included to identify those who are in need of further psychological investigations.

Prematurity is a risk factor for neurodevelopmental disorders.⁴⁰ We adjusted for this confounder as the controls were matched on gestational age at birth. Chromosomal aberrations and syndromes were excluded in this study as they could also be possible confounders of outcome.

The strengths of this study are that up until today it includes the largest study population with the longest follow-up addressing neuro-developmental outcome in patients with EA. All data were prospectively collected with no loss to follow-up and five controls were drawn at random for each individual in the exposed cohort matched on sex, gestational age at birth, birth year and birth county, thus reducing the risk of selection bias.

A limitation of the study is the risk of false negatives as specific criteria were set to avoid misclassification and false positives. Another limitation is the lack of hereditary data that could be associated with the outcomes and no review of medical records was done. The limitations point to the need for future research in these patients including neurodevelopmental evaluations and hereditary data.

CONCLUSION

Individuals with EA have a higher risk of ASD and ID than individuals without the exposure. These results are important when establishing follow-up programs for children with EA to allow timely detection and consequentially an earlier treatment and support. Neurodevelopmental evaluations in preschool children should be included in the national follow-up program for EA so that the need of any additional support and adjustments can be identified before school start.

ABBREVIATIONS

ADHD, Attention deficit/hyperactivity disorder; ASD, Autism spectrum disorders; ATC, Anatomical Therapeutic Chemical code; EA, Esophageal atresia; HRQoL, Health-related quality of life; ICD, International Statistical Classification of Diseases and Related Health Problems; ID, Intellectual disability; NPR, National Patient Register; MBR, Medical Birth Register; PDR, Prescribed Drug Register; CDR, Causes of Death Register; SGA, Small for Gestational age.

ACKNOWLEDGMENTS

We would like to express our gratitude to Fabian Söderdahl, Statisticon AB, for his assistance with the statistical analyses and Sheila Macdonald-Rannström for the final language review.

FUNDING

This study received funding from the HRH Crown Princess Lovisa’s Foundation for Child Health.

Conflict of interest: The authors declare that they have no conflict of interest.

Disclosures: None.

References