Academic Achievements in Adolescents With Congenital Heart Disease: A Total Population‐Based Cohort Study

Abstract

Background

Improved survival in children with congenital heart disease (CHD) has unveiled associated long‐term neurodevelopmental impairments. However, the long‐term effects on academic performance and the influence of sex, family, socioeconomic factors, and surgical era remain understudied, limited by small and incomplete cohorts.

Methods

This total‐population cohort study with nested sibling analysis included all singletons born in Sweden between 1987 and 2005 who had complete socioeconomic data and resided in Sweden at age 16 years. Academic achievement at age 16 was assessed by upper secondary school qualification. Poisson and logistic regression were used to estimate adjusted prevalences, risk ratios, and risk differences. Adjustments were made for year of birth and socioeconomic status. Stratification by noncardiac congenital anomalies, sex, and surgical era was performed.

Results

Of 1 800 477 singletons included, 16 075 (0.9%) had CHD, classified hierarchically as univentricular heart (n=349), severe (n=1939), moderate (n=1764), or mild CHD (n=12 023). After excluding major noncardiac congenital anomalies, the adjusted risk difference for failing to meet upper secondary school qualification compared with children without CHD was 11 (95% CI, 5–17), 7 (95% CI, 4–9), 6 (95% CI, 3–8), and 2 (95% CI, 2–3) additional children per 100 for univentricular heart, severe, moderate, and mild CHD, respectively. Children with both CHD and noncardiac anomalies had poorer academic outcomes than those with CHD alone. Findings were consistent in sibling analyses and over time.

Conclusions

Adolescents with CHD inherently face poorer academic performance, worsening with disease severity and persisting over time.

Nonstandard Abbreviations and Acronyms

GA

gestational age

MCAs

major noncardiac congenital anomalies

USS

upper secondary school

UVH

univentricular heart

Clinical Perspective.

What Is New?

• This nationwide cohort study demonstrates that academic achievement declines with increasing congenital heart disease complexity, with disparities persisting after adjustment for socioeconomic factors at birth and in comparison to siblings.

What Are the Clinical Implications?

• Multidisciplinary follow‐up should extend beyond cardiac care in children with congenital heart disease, to include cognitive and educational support.

• Genetic abnormalities associated with congenital heart disease significantly affect academic outcomes; routine genetic testing should be considered to guide individualized care.

Congenital heart disease (CHD) is the leading major congenital malformation, accounting for nearly one third of all cases. ¹ Advances in pediatric cardiac surgery have been successful, and today, roughly 97% of all children with CHD are expected to survive into adulthood. ² This marks a dramatic improvement, as survival rates in the 1980s were only around 85%. The success is greatly attributed to novel surgical techniques and improved perioperative care. In Sweden, the improved survival coincided with the centralization of pediatric heart surgery in the mid‐1990s. ³ As survival rates in CHD have improved, neurodevelopmental impairments have emerged as an increasingly recognized long‐term consequence. ⁴ This has been highlighted as one of the most significant morbidities affecting their quality of life both in the short and long term. ⁵ , ⁶ , ⁷

The neurodevelopmental challenges in children with CHD are believed to result from a range of factors, including the complexity of the heart disease and concomitant surgery, genetic predisposition, potential complications following surgeries or medications, and socioeconomic factors. ⁴ Brain dysmaturation is present already in fetal life, due to altered fetal circulation and associated fetal cerebral oxygen consumption and has been linked to adverse neurodevelopment. ⁸ , ⁹ Their brain development can be further affected postnatally by altered brain hemodynamics and white matter damage, resulting from both the heart defect itself and the timing, duration, and number of surgeries required for correction. ¹⁰ , ¹¹ , ¹² Consequently, the cumulative disruptive effects on brain maturation may impair neurodevelopment, which, in turn, can affect school performance.

Understanding the impact of CHDs on school achievement is essential, as educational outcomes are well‐established predictors of long‐term quality of life, mortality, and overall health. ¹³ , ¹⁴ Besides reflecting cognitive ability, a key neurodevelopmental domain, grades also reflect broader adaptive functioning, such as the social process of schooling and behavior ¹⁵ and are influenced by socioeconomic factors. ¹⁶ Socioeconomic status also influences neurodevelopmental outcomes in children with CHD ¹⁷ and is considered a risk stratifier ⁴ ; accordingly, it has been emphasized that studies on school performance in children with CHD must account for it. ¹⁸ Previous studies on early childhood academic performance with CHD are limited, and cognitive deficits may not yet be evident. ¹⁹ , ²⁰ Many have been unable to adjust for socioeconomic status and familial confounding or have suffered substantial attrition due to nonresponse, loss to follow‐up, or missing data from private schools. ¹⁸ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ Univentricular heart (UVH), managed by staged palliation culminating in Fontan circulation, represents one of the most complex conditions and is among those most susceptible to neurodevelopmental impairments, ²⁶ which has yet to be delineated in national cohorts of children with CHD. ²¹ , ²² Additionally, it has not been shown whether increased survival rates also improve long‐term neurodevelopmental outcomes. ²

This population‐based cohort study aimed to investigate academic outcomes in adolescents with varying levels of CHD complexity using risk stratification adapted from the 2024 American Heart Association's Scientific Statement on Neurodevelopmental Outcomes.

We assessed school grades and qualification for higher education among 1.8 million Swedish‐born singeltons, adjusting for socioeconomic factors, accounting for familial confounding through sibling comparisons, and stratifying by surgical era to clarify the impact of CHD on educational attainment. In addition, we examined gestational age (GA) and small for gestational age (SGA), as causal pathways for CHD, separately.

METHODS

Data Sources

Using the unique personal identity that is assigned to all Swedish residents, ²⁷ we linked data between multiple registers. Those provided by the Swedish National Board of Health and Welfare were: the Swedish MBR (Medical Birth Register), which has almost complete coverage of all births since 1973 and improved quality since 1982 ²⁸ and the NPR (National Patient Registry), which includes nationwide information on diagnoses for inpatient care since 1987 and hospital outpatient care since 2001. ²⁹ Those provided by the government agency Statistics Sweden were the Total Population Register ³⁰ with information on country of birth, death, and migration; the National School Registers, ³¹ where data on graduation grades are routinely collected in the Registry of Compulsory School Leaving grades; and the Longitudinal Integrated Database for Health Insurance and Labor Market Studies, ³² providing information on parental education and income. The study was conducted in accordance with the Helsinki Declaration and approved by the Regional Ethical Review Board in Stockholm, Sweden (DNR: 2020‐05516, 2021‐05958‐02, 2022‐05648‐02 and 2024‐00060‐01). According to current Swedish regulations, no informed consent is required for research using national registry data. The data used in this study were obtained from Swedish national registers and cannot be made publicly available due to restrictions under the European General Data Protection Regulation and Swedish data protection legislation. Analytical methods, including statistical code, will be shared upon reasonable request to the corresponding author.

Study Population

This population‐based cohort study included 1 938 397 births recorded in the MBR from January 1, 1987 to December 31, 2005. A flow chart for inclusion and exclusion is shown in Figure S1.

Congenital Heart Disease

Diagnoses of CHD and compatible surgery or intervention procedure codes were retrieved from the NPR using the International Classification of Diseases, Ninth Revision (ICD‐9) and Tenth Revision (ICD‐10) and the Classification of surgical procedures (K06 and KVÅ/KKÅ). ³³ Following the New England Regional Infant Cardiac Program, ³⁴ a coding hierarchy was used to classify CHD, ensuring each lesion complexity was assigned to 1 and only 1 of 69 designated hierarchical numbers. Although not exhaustive, the list of included diagnoses represents the most common types of CHD. These diagnoses were further grouped by complexity, taking into account timing of surgery (in infancy, childhood, or no surgery at all), considering the risk of adverse neurodevelopmental outcomes as outlined in the 2024 American Heart Association's Scientific Statement on Neurodevelopmental Outcomes. ⁴ We were also able to create 1 group exclusively for the UVH.

Diagnoses of CHD were obtained before the expected attainment of final grades from compulsory school at age 16 years. For the UVH group, a diagnosis aligned with single ventricle and a surgical procedure code consistent with at least step 2 of 3 in Fontan palliation was required. ³⁵ The surgical procedure code alone for UVH was also accepted, but the primary diagnoses of the cases were reviewed separately to ensure accuracy. For the severe CHD group, a surgical procedure code matching the CHD leading to a biventricular repair or the procedure code of having been treated with cardiopulmonary bypass was required within the first year of life. For children classified as moderate CHD, a surgical procedure code aligned with the CHD was required before the expected attainment of final grades. For the mild unoperated group (mild CHD), all surgical procedure codes for heart surgery (Chapter 3 in K06 and Chapter F in KVÅ/KKÅ ³³ ) were used to omit surgical cases, ensuring only nonsurgical cases were included. A list of all included CHD diagnoses, organized hierarchically and grouped by complexity based on compatible surgical or interventional procedure codes, is provided in Table S1.

Children with patent ductus arteriosus associated with prematurity or acquired mitral insufficiency due to rheumatic disease were identified using ICD‐9 or ICD‐10 and removed from the CHD group.

For certain diagnoses typically requiring surgery within the first year of life (including common arterial trunk, aortopulmonary septal defect, double outlet right ventricle, atrioventricular septal defect, total anomalous pulmonary venous drainage, pulmonary atresia with intact ventricular septum, interrupted aortic arch, and tetralogy of Fallot) only those children who actually underwent cardiac surgery during the first year were retained in the severe CHD group. Children with these diagnoses who did not undergo surgery were removed from the cohort with CHD, as the absence of a matching procedure code may indicate miscoding of the diagnosis, since surgical intervention within the first year of life is the expected management.

Noncardiac congenital anomalies (CAs) were identified using ICD‐10 (Q00‐Q99) and ICD‐9 (740–759). Major noncardiac CAs (MCAs) excluded CHD ICD‐10 (Q20–Q26) and ICD‐9 (745–747) and minor noncardiac CAs as defined by the Swedish National Board of Health and Welfare ³³ , ³⁶ and presented in Table S2.

As MCAs are associated with a high prevalence of neurodevelopmental impairment that may confound CHD‐specific effects, primary analyses were conducted in children without MCAs. However, all analyses were repeated with MCAs included to ensure transparency, and results are presented in parallel throughout.

The Swedish Compulsory School System

In Sweden, compulsory schooling lasts 9 years and typically begins when the child turns 7. Private schools are relatively few and almost all operate within the public system, receiving public funding and charging no tuition fees. All Swedish schools, whether public or private, are required to report grades to the Swedish National Agency for Education. Home schooling is not permitted under Swedish law.

School grades in Sweden reflect academic performance but are also affected by school attendance, as sufficient attendance is required for teachers to assess and grade the student. Academic performance is assessed partly through national standardized tests, administered on the same day across the entire country. These tests serve as a guiding tool for teachers to support consistent and comparable grading practices nationwide.

Children with intellectual disabilities are offered special education and typically attend separate special education schools. These schools are the only school form that does not report grades to the national registry, as students follow an adapted curriculum. ³¹

The Swedish school system and the nationally standardized grading scale is described in detail in Figure S2. The grading system was slightly modified in 2012. Both systems were based on criterion‐referenced grades, with the same distinction between pass and no pass, where a pass corresponded to 10 grade points and the maximum grade per subject to 20 grade points. With an initial maximum of 16 subjects, this yielded a maximum total grade sum of 320 in the earlier system, which increased to 340 in the latter, provided an additional grade in a modern language was also achieved. Eligibility criteria for university studies were also similar across both systems.

Main Outcome Measures

Qualification for upper secondary school (USS) in children by age 16 years without MCAs was our primary outcome, which consists of a composite of grades in English, Swedish, and mathematics, along with an additional 5 subjects from the ninth year of compulsory school in the Swedish school system.

The outcome variable comprised 3 categories: (1) qualification yes or (2) no, or (3) if the student had not completed the last year of compulsory school, leading to nonreported grades (neither pass nor no pass in any subject). The latter is applicable for school dropouts or students with intellectual disabilities requiring special schooling. Not having grades reported in the system has previously been used for imputation as having grades within the lowest first percentile. ³⁷

In a subanalysis, we evaluated the risk of receiving only a passing grade rather than a higher grade in the core subjects and the total grade point, which comprises the sum of all grades, ³¹ to gain a more nuanced assessment of the child's scholastic performance. We imputed missing grades as no grades. ³⁷

Covariates

We evaluated 2 main categories of potential factors associated with academic performance:

1. Child demographic characteristics such as sex ³⁸ and year of birth. ³⁹

2. Sociodemographic characteristics ¹⁷ , ⁴⁰ (highest education of parents by the time of birth, country of birth for parents, disposable income/family member at time of birth,* maternal and paternal age by birth).

*Individualized disposable family income reflects the financial situation of a household after accounting for all sources of income (including social benefits) for a family, deduction of all outgoing expenses (including taxes), and adjustment for family size. It is updated yearly by Statistics Sweden and is based on mandatory reporting. We extracted the individualized family income at the time of the child's birth and, to avoid misclassification bias because of inflation, coded this into quintiles in relation to year of ascertainment.

Additional clinical risk factors attributed to CHD and possibly affecting neurodevelopmental outcome include noncardiac CAs, ⁴¹ GA, ²¹ and relative birth weight ¹⁷ (calculated as the centile of birth weight within the given gestational age ⁴² ). However, to avoid overadjusting for GA and SGA, which are causal pathways for CHD, they were analyzed separately in relation to their impact on the outcome in children with CHD. In the current study, we used the more stringent definition of SGA below the third percentile to better capture pathological fetal growth restriction and reduce misclassification of constitutionally small infants. ⁴³

A secondary analysis was performed to elaborate the risk factor of being born before or after the centralization of pediatric heart surgery (1992–1996), ³ stratified on period by birth year. This was merely exploratory since survival significantly increased after centralization. ³

Exclusion

We excluded multiple births (n=53 197), children who emigrated (n=37 854), or children who died before graduation at the age of 16 years (n=10 845). There were no missing data on children's or parents' personal identification numbers, and no missing information on sex. Missing socioeconomic data (n=36 025) led to excluding cases that could not be adjusted. Additionally, all children with pulmonary atresia with ventricular septal defect and major aortopulmonary collateral arteries (n=19) were excluded, as they represent a distinct group with high morbidity but with too few cases.

Statistical Analysis

All analyses were conducted by using SAS (version 9.4). One author (M.L.) had full access to all study data and takes responsibility for their integrity and analysis. This article conforms to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines for observational studies. We used multivariable Poisson regression models with robust variance to estimate adjusted risk ratio (RR) (with 95% CI) for each academic outcome by the status of CHD. To estimate risk differences, an identity link was added to the model. Risk differences are presented in percentages (%). Total grade sum was evaluated by multiple regression models within the framework of general linear models. In addition to the status of CHD, the models included adjustments for covariates such as year of birth, parents' highest education at the time of birth, country of birth for parents, family income (disposable income/family member), and maternal and paternal age at birth. Major and minor congenital anomalies were either excluded or retained in the analysis. Children were stratified by sex, GA, and relative birth weight.

Sensitivity Analysis

To evaluate confounding bias from school dropouts and children in special schools unable to receive final grades, we investigated the effect of adjusting the inclusion of children without information on grades by imputing them as not having passed grades in the subjects with missing information, grade sum=0. ³⁷ The group of children with CHD had proportionally fewer grades compared with the total population.

Finally, shared family factors, as well as unmeasured factors (eg, parental genetic or unmeasured social factors), were considered using sibling analysis, because siblings share 50% of genetics. ⁴⁴ In a subset of families with >1 child, we matched each child not qualified for higher education (cases) with ≥1 full siblings qualified for higher education (controls). To compare the risk of non‐eligibility in children with congenital heart disease with that of their qualified siblings, we estimated odds ratios (ORs) in conditional logistic regression models matched on family identification number, adjusted for nonshared confounding characteristics.

RESULTS

Among 1 938 397 liveborn children in Sweden between 1987 and 2005, 1 800 477 were retained in the current study, identifying 16 075 individuals with congenital heart disease. Of these 349 (2.2%) had UVH, 1939 (12.1%) had severe CHD, 1764 (11.0%) had moderate CHD, and 12 023 (74.8%) had mild, nonsurgical CHD. Figure S1 shows the flow chart of inclusion/exclusion. The sum of missing information accounted for 2.0% and was due to incomplete information on socioeconomic factors.

Among children with CHD, 32.2% had a concomitant MCA and 39.8% had any CA, compared with the total population excluding CHD, where 5.5% had an MCA and 10.2% had any CA, as described together with baseline characteristics in Table 1. In total, 254 670 children were identified in the subset of families with at least 1 sibling nonqualified for USS and included in the sensitivity analysis.

Table 1.

Characteristics of Parents and Their Singletons Born in Sweden, 1987 to 2005, With or Without CHD, and Residing in Sweden at Age 16 Years

Characteristics	Total	No CHD	Complexity of CHD
			UVH	Severe CHD	Moderate CHD	Mild CHD
Total	1 800 477	1 784 402	349	1939	1764	12 023
Mothers
Age at delivery, y±SD	29.4 (5.1)	29.4 (5.1)	30.2 (5.0)	30.3 (5.3)	29.9 (5.4)	30.1 (5.2)
Country of birth
Sweden	1 541 466 (85.6)	1 527 777 (85.6)	300 (86.0)	1634 (84.3)	1452 (82.3)	10 303 (85.7)
Nordic (except Sweden)*	50 232 (2.8)	49 837 (2.8)	7 (2.0)	61 (3.2)	51 (2.9)	276 (2.3)
Other European	29 035 (1.6)	28 785 (1.6)	4 (1.1)	39 (2.0)	32 (1.8)	175 (1.5)
Outside Europe	179 744 (10.0)	178 003 (10.0)	38 (10.9)	205 (10.6)	229 (12.1)	1269 (10.6)
Level of highest education†
9 y	281 439 (15.6)	278 999 (15.6)	48 (13.8)	298 (15.4)	326 (18.5)	1768 (14.7)
10–12 y	951 559 (52.9)	943 516 (52.9)	165 (47.3)	1005 (51.8)	889 (50.4)	6094 (50.7)
>12 y	567 369 (31.5)	561 887 (31.5)	136 (39.0)	636 (32.8)	549 (31.1)	4161 (34.6)
Fathers
Age at time of birth, y±SD	32.2 (6.0)	32.2 (6.0)	33.0 (6.4)	33.1 (6.4)	32.7 (6.4)	32.9 (6.1)
Country of birth
Sweden	1 529 779 (85.0)	1 516 229 (85.0)	293 (84.0)	1635 (84.3)	1435 (81.4)	10 187 (84.7)
Nordic (except Sweden)*	46 125 (2.6)	45 754 (2.6)	10 (2.9)	34 (1.8)	47 (2.7)	280 (2.3)
Other European	29 296 (1.6)	29 062 (1.6)	3 (0.9)	39 (2.0)	24 (1.4)	168 (1.4)
Outside Europe	195 277 (10.9)	193 357 (10.8)	43 (12.3)	231 (11.9)	258 (14.6)	1388 (11.5)
Level of highest education†
≤9 y	322 059 (17.9)	319 360 (17.9)	59 (16.9)	333 (17.1)	347 (19.7)	1960 (16.3)
10–12 y	953 908 (53.0)	945 395 (53.0)	169 (48.4)	1014 (52.3)	901 (51.1)	6429 (53.5)
>12 y	524 510 (29.1)	519 647 (29.1)	121 (34.7)	592 (30.5)	516 (29.2)	3634 (30.2)
Socioeconomic status
Total family income/individual‡
Percentile 0–20	309 019 (17.2)	306 259 (17.2)	58 (16.6)	335 (17.3)	337 (19.1)	2030 (16.9)
Percentile 20–40	345 210 (19.2)	342 078 (19.2)	59 (16.9)	384 (19.8)	333 (18.9)	2356 (19.6)
Percentile 40–60	367 768 (20.4)	364 453 (20.4)	71 (20.3)	408 (21.0)	358 (20.3)	2478 (20.6)
Percentile 60–80	384 026 (21.3)	380 571 (21.3)	87 (24.9)	374 (19.3)	345 (19.6)	2649 (22.0)
Percentile 80–100	394 454 (21.9)	391 041 (21.9)	74 (21.2)	438 (22.6)	391 (22.2)	2510 (20.9)
Children
Sex
Female	875 117 (48.6)	867 310 (48.6)	138 (39.5)	829 (42.8)	924 (52.4)	5916 (49.2)
Male	925 360 (51.4)	917 092 (51.4)	211 (60.5)	1110 (57.3)	840 (47.6)	6107 (50.8)
Small for gestational age 42
<Third percentile	42 181 (2.3)	41 223 (2.3)	12 (3.4)	143 (7.4)	147 (8.3)	656 (5.5)
Missing	6378 (0.4)	6221 (0.3)	7 (2.0)	43 (2.0)	20 (1.0)	87 (0.7)
Gestational age
>36 wk	1 711 325 (95.0)	1 697 498 (95.1)	318 (91.1)	1721 (88.8)	1451 (82.2)	10 337 (86.0)
<37 wk	87 045 (4.8)	84 824 (4.8)	31 (8.9)	209 (10.8)	310 (17.6)	1671 (13.9)
Missing	2107 (0.1)	2080 (0.1)	0 (0)	9 (0.5)	3 (0.2)	15 (0.1)
Congenital anomalies
No	1 614 313 (89.7)	1 603 019 (89.8)	170 (48.7)	1023 (52.8)	1118 (63.4)	8983 (74.7)
Major noncardiac	102 327 (5.7)	98 459 (5.5)	153 (43.8)	828 (42.7)	546 (31.0)	2341 (19.5)
Noncardiac	186 164 (10.3)	181 383 (10.2)	179 (51.3)	916 (47.2)	646 (36.6)	3040 (25.3)
Follow‐up (major noncardiac congenital anomaly excluded)
Not in school registry (%)		2.2	9.2	6.0	6.3	3.8
Approved mathematics (%)		90.6	81.1	83.2	84.6	87.8
Approved Swedish (%)		93.4	85.2	87.4	88.0	91.2
Approved English (%)		92.7	84.2	86.9	86.4	90.6

Values are numbers (percentages) unless stated otherwise.

CHD indicates congenital heart disease; and UVH, univentricular heart disease.

^*Finland, Norway, Denmark, and Iceland.

^† Highest educational level attained at time of childbirth in number of years of completed education of both parents into; up to 9 y (primary school), 10 to 12 y (secondary school), and >12 y (higher education or university). See Figure S2 for an overview of the Swedish school system.

^‡ Individualized disposable family income reflects the financial situation of a household after accounting for all sources of income (including social benefits) for a family, deduction of all outgoing expenses (including taxes), and adjustment for family size. It is updated yearly by Statistics Sweden and is based on mandatory reporting. We extracted the individualized family income at the time of the child's birth and, to avoid misclassification bias because of inflation, coded this into quintiles in relation to year of ascertainment.

CHD and Qualification for USS

Overall, children with CHD had worse academic performance than did the children without CHD, who served as the reference group, a difference that was stable after adjusting for potential confounders (sex, birth year, maternal age, paternal age, maternal birth country, maternal education, paternal education, and disposable income/family member). The primary outcome measure, qualification for USS, in children without noncardiac MCAs, under the assumption that those without any grade were deemed ineligible nonqualified, demonstrated a linear association with the complexity of CHD (Figure 1). In the entire study population, also including children with MCA apart from CHD, the greatest disparity was shown for children with a severe CHD (RR, 3.37 [95% CI, 3.11–3.50]) of not being qualified compared with the reference group (Figure S3).

Figure 1. The risk of not being qualified for USS at 16 years of age in relation to the complexity of CHD, MCA excluded.

Data were analyzed using multivariate Poisson regression models, adjusted for sex, birth y, maternal age, paternal age, maternal birth country, paternal birth country, maternal education, paternal education, and disposable income/family member. Data are presented as RR with 95% CIs and RD as a percentage (%). Missing grades were imputed as not passing. CHD indicates congenital heart disease; MCA, major noncardiac congenital anomaly; RD, risk differences; Ref, reference group; RR, relative risk; and USS, upper secondary school.

Boys and girls without noncardiac MCA demonstrated comparable qualification for USS (Table S3). Qualification for USS showed a similar significant trend with nonimputed grades in adjusted and nonadjusted models (Figure S4; Table S3).

As minor CA did not affect the results, they were retained in the following analysis to improve power (Table S3).

CHD, Grades in Core Subjects and Total Grade Sum

The complexity of CHD (noncardiac MCA excluded) was associated with the risk of failing core subjects in ninth grade, with an equally elevated risk across all 3 core subjects (Table 1). The probability of achieving a higher grade than pass in the core subject was also hampered in all children with CHD after adjustments, except for the moderate cases, and was borderline significant for UVH cases in mathematics (Table S4).

Elucidating the broader impact of CHD on the total grade sum in ninth grade, we observed a consistent trend similar to that seen in core subjects. When excluding MCA and stratifying by sex, total grade sum was affected by both the complexity of CHD and sex, where boys with UVH had an average corresponding to just above pass grades in all subjects, but girls with UVH had somewhat higher total grade sum, as shown in Table 2.

Table 2.

Total Grade Sum for Children With or Without CHD, MCA Excluded

Sex	CHD status	Total crude	Total adjusted*	P value for trend
All
	No CHD	211 (211–211)	211 (211–211)
	UVH	178 (168–187)	174 (165–183)
	Severe CHD	193 (189–198)	193 (189–197)	P<0.001
	Moderate CHD	197 (193–201)	196 (192–199)
	Mild CHD	208 (207–210)	204 (202–205)
Male sex
	No CHD	199 (199–199)	199 (199–199)
	UVH	178 (165–191)	168 (157–180)
	Severe CHD	189 (184–194)	186 (182–191)	P<0.001
	Moderate CHD	186 (180–192)	186 (180–191)
	Mild CHD	196 (194–198)	192 (190–194)
Female sex
	No CHD	223 (223–223)	223 (223–223)
	UVH	177 (162–192)	178 (164–191)
	Severe CHD	200 (194–207)	197 (191–203)	P<0.001
	Moderate CHD	205 (200–211)	206 (202–211)
	Mild CHD	221 (219–223)	216 (214–217)

Data are presented as the total grade sum with a 95% CI. Imputation of no grade in a subject as 0 grade points. All individuals and stratification by sex are presented. CHD indicates congenital heart disease; MCA, major noncardiac congenital anomaly; and UVH, univentricular heart disease.

^*Adjusted for sex, birth year, maternal age, paternal age, maternal birth country, maternal education, paternal education, and disposable income/family member.

Academic Performance by SGA and GA

SGA and prematurity, individually, nearly doubled the risk of being nonqualified for USS in all children with CHD without noncardiac MCA, with RR 2.34 (95% CI, 2.04–2.69) and 1.81 (95% CI, 1.64–1.99), respectively. However, when subgrouping children in complexity of CHD, significant differences remained for moderate and mild cases without MCAs, compared with children with the same complexities who were not premature or SGA. A nonsignificant additional risk of being SGA or preterm was observed for UVH and severe CHD; however, these cases were few, which may have led to underpowering in these groups (Figure 2).

Figure 2. Additive relative risk of being SGA (<3 SD) or preterm (<37 GA) in relation to the complexity of CHD and nonqualification for USS, MCA excluded.

Data were analyzed using multivariate Poisson regression models, adjusted for sex, birth y, maternal age, paternal age, maternal birth country, paternal birth country, maternal education, paternal education, and disposable income/family member. Children with an MCA were excluded. Data are presented as RR with 95% CIs and RD as a percentage (%). Missing grades were imputed as not passing. AGA indicates appropriate for gestational age; CHD, congenital heart disease; GA, gestational age; MCA, major noncardiac congenital anomaly; Ref, reference; RR, relative risk; SGA, small for gestational age ⁴² ; USS, upper secondary school; and UVH, univentricular heart disease.

Qualification for USS by Period

In our exploratory analysis, stratifying by surgical eras based on birth year, before (1987–1991), during (1992–1996), and after the centralization of pediatric heart surgery (1997–2005), did not reveal any change in the prevalence of qualification for higher education over time across all complexities of CHD in children without noncardiac MCA (Figure 3).

Figure 3. Relative risk of not being qualified for USS by birth cohorts in children with CHD, MCA excluded.

The birth cohort from 1987 to 1991 represents the period before, 1992 to 1996 during, and 1997 to 2001 and 2002 to 2005 after the centralization of pediatric heart surgery in Sweden. Data are presented as RR with a 95% CI. Missing grades were imputed as not passing. No CHD serves as the reference. Adjusted for sex, birth year, maternal age, paternal age, maternal birth country, paternal birth country, maternal education, paternal education, and disposable income/family member. CHD indicates congenital heart disease; MCA, major noncardiac congenital anomaly; RR, relative risk; USS, upper secondary school; and UVH, univentricular heart disease.

Academic Performance in Relation to Siblings

Sibling comparisons revealed the same risk of not being qualified for USS across the complexities of CHD, as when compared with the general population, after excluding MCA. For the entire group, including all with CAs, differences between siblings were somewhat more pronounced than in the general population (Table S5). Stratification by sex revealed a comparable risk of being nonqualified across all levels of complexity of CHD. However, the analysis was limited by wide CIs, attributed to attrition, thereby reducing statistical power.

DISCUSSION

In this comprehensive nationwide cohort of 1.8 million children, all CHD was associated with poorer academic performance in adolescence. Notably, the extent of academic impairment increased with the complexity of the CHD. Outcomes worsened with MCAs, prematurity, and SGA. The findings remained consistent across sibling comparisons by sex and birth cohorts.

Impaired cognitive outcomes are the most common self‐reported disabilities in adults with CHD, ⁷ affecting their socioeconomic status ⁴⁵ and increasing mortality. ⁴⁶ Unlike previous studies, which are numerous and often conducted in childhood, showing impaired intelligence or executive functions through single tests, ²¹ , ⁴⁷ , ⁴⁸ our study demonstrates the combined effect of multiple neurodevelopmental factors on academic grades in adolescence which may better reflect long‐term outcomes. ⁴⁹ Although qualification for USS is not a neurodevelopmental outcome per se, it captures the cumulative real‐world expression of key neurodevelopmental domains and represents a critical transition point to higher education. Prior research on academic performance has yielded inconsistent results. ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ Studies in early childhood often show less difference in academic achievement in children with CHD compared with healthy peers, ²² potentially overestimating their later performance. ⁵⁰ , ⁵¹ Long‐term data from large groups indicate lower educational attainment, with 1 study showing a declining trend over the past 2 decades. ²¹ However, these studies either lacked socioeconomic adjustments or failed to adequately account for the need for surgery. ¹⁹ , ²¹ , ²²

Our findings reveal that poorer academic outcomes were consistent over time across severity groups, indicating that increased survival has not led to improved neurodevelopment. The current study also broadened the insight into academic outcomes, looking not only at core subjects but at the composite grade of all subjects, which has previously been unrecognized. The trend of lower achievement in core and all subjects is consistent with increasing CHD complexity. Although socioeconomic factors are recognized as major determinants of neurodevelopmental outcomes in children with CHD, ¹⁷ our results remained largely consistent even after accounting for socioeconomic conditions at birth and shared familial factors in sibling analysis. This resilience may partly reflect the mitigating effect of Sweden's universal, publicly funded health care and education systems. However, because socioeconomic conditions were captured only at birth, the long‐term impact on socioeconomic trajectories over the follow‐up period to age 16 years may not be fully accounted for.

We also found that children with moderate and mild nonsurgical CHD had slightly impaired academic performance, in contrast with some previous studies on neurodevelopmental outcomes. ⁵² , ⁵³ Growing knowledge of genetic anomalies linked to CHD and neurological disabilities, together with the known hemodynamic effects of CHD during fetal life, may explain these findings. ¹⁰ , ²⁶ Our results demonstrating that prematurity adds additional negative effects on educational outcomes align with recent research, ²¹ suggesting a potential “double‐hit” effect on neurodevelopment as previously outlined. ⁵⁴ , ⁵⁵ In the general population, late preterm birth is associated with worse neurodevelopmental outcomes, ⁵⁶ though this is not mirrored in their academic performance. ³⁷

We also find that SGA affects academic achievement, as previously shown to worsen neurodevelopment. ¹⁷ The 2024 American Heart Association's Scientific Statement on Neurodevelopmental Outcomes identifies growth failure in infancy and toddlerhood as a risk factor for impaired neurodevelopment, yet it includes birth weight <10th percentile among its defining criteria. Our findings, captured by the more stringent definition of SGA (birth weight <third percentile), warrant distinct emphasis, as it has been shown to identify true fetal growth restriction more accurately and is more strongly associated with adverse neurodevelopmental outcomes than the 10th percentile threshold. ⁵⁷

Recognition of the reduced likelihood of higher educational attainment among children with CHD, particularly those with more complex conditions, underscores the need for strategies to mitigate these disparities. Recently, a neurodevelopmental program in infancy has shown short‐term cognitive gains. ⁵⁸ In contrast, an intervention aimed at improving cognition in childhood has not yielded similar benefits in children with CHD. ⁵⁹ Moreover, screening for neurodevelopmental comorbidities in children with CHD could be challenging, as they may not fulfill criteria for any single diagnosis, yet still experience a range of difficulties across the neurodevelopmental domains. ⁴ Acknowledging these challenges is essential to ensure that children with CHD receive appropriate support from child psychiatric, educational, and social services. ⁶⁰ Although the overall burden of neurodevelopmental disorders in the CHD population remains to be fully elucidated, emerging evidence indicates a significantly increased prevalence. ⁵ The current data justify the change in obstetrical practice with increasing gestational age at delivery. ³⁹ For clinicians, these findings provide additional insight to support antenatal counseling. As prenatal detection rates for all complex CHD increase, and for UVH now approach 100%, ⁶¹ parents are increasingly faced with complex decisions regarding pregnancy continuation, underscoring the need for robust evidence on long‐term neurodevelopmental outcomes. Moreover, children with CHD and a concurrent MCA demonstrated significantly poorer academic outcomes than those with CHD alone in our current study. This is likely explained by CHD‐associated MCAs such as Down syndrome and 22q11 deletion syndrome, where the underlying genetic aberrancies independently influence neurodevelopmental trajectories. Accordingly, sibling comparisons showed more pronounced differences when MCA was included, as unaffected siblings do not carry these genetic aberrancies. These findings underscore the importance of genetic testing and counseling in this population.

Strengths and Limitations of This Study

A major strength of this study is its large, population‐based design set in a contemporary context. Additionally, the Swedish population is diverse, enhancing generalizability. ⁶² The Swedish registers provide extensive, high‐quality data, allowing us to examine nearly all Swedish children with CHD and their academic achievements, while adjusting for socioeconomic factors. Our imputation of grades uncovered greater disparities between children with CHD and the overall population. Significant differences remained in both unadjusted models and those incorporating nonimputed grades, thereby establishing credibility. Our inclusion criteria, based on diagnosis, surgical procedure codes, and timing of surgery, may have misclassified some children with CHD as controls, potentially underestimating associations. ⁶³ Our proportion of children with CHD is well in line with reported prevalences, ¹ , ² although recognizing an increasing trend among mild CHD cases, such as patent ductus of prematurity, which was not considered a CHD in the current study. The number of severe cases in our cohort is most likely underestimated. We identified 12.6 per 10 000 children alive by age 16 years, whereas a US‐based study reported a birth prevalence of 17.3 per 10 000 live births, corresponding to ∼14.3 per 10 000 at 1 year of age after accounting for first‐year mortality. ⁶⁴ Our lower estimate likely reflects a combination of mortality between infancy and by the age 16, differences in the definition of severe CHD, and our requirement for concordance between diagnosis and procedure codes, which increases specificity at the expense of sensitivity. Nevertheless, our reported number of UVH cases aligns with a recent publication on long‐term survival after univentricular palliation in Sweden. ³⁵ Moreover, 2.0% of the population were excluded from the study cohort due to missing socioeconomic data, and 5.1% were excluded due to multiparty, emigration, or death.

During our years of inclusion, the Swedish school system underwent one change in grading, introducing on more grade on the scale in 2012. We adjusted the analysis for year of birth, but the change in the grading system remains a limitation. Moreover, the exact number of children requiring special education in the current population remains unknown. Another limitation is that the number of early‐term deliveries declined in Sweden between 2006 and 2014, ³⁹ which could result in improved school performance in the current cohort, because even early‐term delivery has been linked to adverse neurodevelopmental outcomes, especially in children with UVH. ⁵⁴ , ⁵⁵

CONCLUSIONS

This extensive population‐based study found that there is an inherent risk of adverse academic outcomes in adolescence in all children with CHD that increases with complexity. Fortunately, we found that most children with CHD will be qualified for higher education. Still, health care personnel and policymakers need to be aware of the children's challenges. Further studies should focus on efficient interventions to improve neurodevelopmental outcomes in children with CHD, particularly since they comprise a significant and growing population.

Sources of Funding

Sofia Ekmark‐Sergel received funding from the Swedish Heartchild Foundation, Veronica Siljehav from the Donation Fund for Research on Neurodevelopmental Disorders at Karolinska University Hospital, Gunnar Bergman from the Swedish Heart Lung Foundation, and Ulrika Ådén from the Regional ALF Grants Region Östergötland. The funders had no role in any aspect of the study, and all authors acted independently from the study sponsors.

Disclosures

None.

Supporting information

Tables S1–S5

Figures S1–S4

STROBE checklist