Congenital heart defects in children born after assisted reproductive technology: a CoNARTaS study

Abstract

Background and Aims

Children born after assisted reproductive technology (ART) have worse perinatal outcomes compared with spontaneously conceived children. This study investigates whether children conceived after ART have a higher risk of congenital heart defects (CHDs) compared with children born after spontaneous conception (SC).

Methods

All 7 747 637 liveborn children in Denmark (1994–2014), Finland (1990–2014), Norway (1984–2015), and Sweden (1987–2015), where 171 735 children were conceived after ART, were included. National ART and medical birth registry data were cross-linked with data from other health and population registries. Outcomes were major CHDs, severe CHDs, 6 hierarchical CHD lesion groups, and 10 selected major CHDs, diagnosed prenatally or up to 1 year of age (Denmark, Finland, and Sweden) and prenatally or at birth (Norway). The association between ART and CHDs was assessed with multivariable logistic regression analysis, with adjustment for available confounders.

Results

Major CHDs were detected in 3159 children born after ART (1.84%) and in 86 824 children born after SC [1.15%; adjusted odds ratio (AOR) 1.36; 95% confidence interval (CI) 1.31–1.41]. Risk was highest in multiples, regardless of conception method. Severe CHDs were detected in 594 children born after ART (0.35%) and in 19 375 children born after SC (0.26%; AOR 1.30; 95% CI 1.20–1.42). Risk was similar between ICSI and IVF and between frozen and fresh embryo transfer.

Conclusions

Assisted reproductive technology–conceived children have a higher prevalence of major CHDs, being rare, but severe conditions. The absolute risks are, however, modest and partly associated with multiple pregnancies, more prevalent in ART.

Structured Graphical Abstract

The main findings were that assisted reproductive technology (ART) was associated with an increased risk of major congenital heart defects (CHDs) as well as severe CHDs in liveborn children with follow-up to 1 year of age, compared with spontaneous conception (SC). Children born from a multifetal pregnancy had the highest risk of CHDs, but ART was also associated with an increased risk in singletons. No significant difference was found between singletons born after intracytoplasmic sperm injection (ICSI) and in vitro fertilization (IVF) or between fresh and frozen embryo transfer. AOR, adjusted odds ratio; ASD, atrial septal defect; CI, confidence interval; FET, frozen embryo transfer; HLHS, hypoplastic left heart syndrome; VSD, ventricular septal defect.

Congenital heart defects, Assisted reproductive technology, Frozen embryo transfer, In vitro fertilization, Intracytoplasmic sperm injection, Registry-based study

See the editorial comment for this article ‘Assisted reproductive technology and heart defects: what’s real and what’s not?’, by N. Auger et al., https://doi.org/10.1093/eurheartj/ehae549.

Introduction

The field of reproductive medicine is growing due to advancements in assisted reproductive technology (ART).¹ More than 10 million children are so far conceived through ART worldwide, and currently, 3.0% of children in Europe and 2.3% in the USA are born after ART.^2–4

Health outcomes for children born after ART continue to be in focus due to the widespread use of ART, the increasing number of children born after ART, and the fast development of new ART procedures. Many systematic reviews, meta-analyses, and large observational studies show an association between ART and low birth weight (LBW) and preterm birth (PTB). Although multiple births are the most important cause of the increased risk of PTB and LBW in ART-conceived children compared with children born after spontaneous conception, risk of these outcomes is also higher in ART-conceived singletons.^5–8 Several meta-analyses and original studies have found that birth defects are more common in children born after ART compared with children born after spontaneous conception. Estimates of excess risk range between 30% and 70%.^7–10 A systematic review, including 29 studies, by Qin et al.⁷ reported birth defects in 5.7% [95% confidence interval (CI) 4.7%–6.9%] of singletons born after in vitro fertilization (IVF)/intracytoplasmic sperm injection (ICSI) and in 3.9% (95% CI 3.1%–4.8%) of singletons born after spontaneous conception. Most studies show no difference in the frequency of birth defects in children born after IVF vs. ICSI, nor in children born after fresh vs. frozen-thawed embryo transfer (FET).^10,11 The increased rate of birth defects found among children born after ART could partly be explained by parental subfertility according to a recent study from Australia.¹²

Congenital heart defects (CHDs) refer to structural anomalies of the heart and the intrathoracic vessels present during pregnancy or at birth.¹³ Congenital heart defects are the most commonly occurring birth defects, accounting for ∼50% of all major birth defects, affecting ∼1%–2% of children in the general population.^14–17 Although many CHDs are recorded at birth and a few later in life, the true incidence of CHDs remains largely unknown due to lack of identification in pregnancies ending in miscarriages, terminations, or stillbirths.^13,18,19 Heart defects are a major paediatric health concern and remain the leading cause of mortality from birth defects.²⁰ Several systematic reviews and cohort studies have found an increased risk of CHDs in children born after ART.^21–23 A recent review including 41 cohort and case–control studies with 25 856 children born after ART showed an increased risk of CHDs in the ART-conceived group as compared with spontaneous conception [SC; pooled odds ratio (OR) 1.45; 95% CI 1.20–1.76].²¹ Congenital heart defects are more common in twins compared with singletons,^24,25 and a recent study found that most of the association between ART and CHD was mediated by twinning.²⁶ For specific CHDs, conflicting results have been reported.^21,27

Using nationwide data from four Nordic countries, we assessed the risk of major CHDs in ART-conceived liveborn children compared with children born after spontaneous conception. We further explored if risk of any specific CHD was increased in children born after ART and if specific assisted reproductive techniques were associated with CHDs.

Methods

Data sources

The Committee of Nordic ART and Safety (CoNARTaS) was established in 2008 to evaluate short- and long-term health consequences of ART in children and their mothers.²⁸ The unique personal identity number, assigned to all residents in the Nordic countries, enabled individual-level data linkage between children and their mothers and between different registries.²⁹ Data from national ART registries, medical birth registries (MBRs), national patient registries (NPRs), cause of death registries, and population registries were cross-linked. Data for this study were obtained from Denmark (1994–2014), Finland (1990–2014), Norway (1984–2015), and Sweden (1987–2015). Due to incompleteness of the Swedish ICD-8 codes for 1985 and 1986, we chose to exclude these 2 years for Sweden. Details on our cohort and the registries used are given in Supplementary data online, Table S1.

Study population

Inclusion criteria were all liveborn singletons, twins, and higher-order multiples born after ART and SC (i.e. conception without ART) during the study period. Assisted reproductive technology is defined according to Zegers-Hochschild et al.,³⁰ i.e. ‘all interventions that include the in vitro handling of both human oocytes and sperm or of embryos for the purpose of reproduction’.

Stillbirths were excluded due to low data quality on birth defects in these pregnancies. Information on terminations of pregnancies due to birth defects was not available.

Outcome variables

We defined children with CHDs as having a CHD diagnosis at birth and up to 1 year of age, in the MBR, NPR, or cause of death registry. For Norway, follow-up stopped at birth. All diagnoses were coded according to the International Classification of Diseases (ICD), Eighth Revision (ICD-8); Ninth Revision (ICD-9); and International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10).³¹

The primary outcome was major CHDs diagnosed up to 1 year of age. Major CHDs were defined according to the European Concerted Action on Congenital Anomalies and Twins (EUROCAT) as ICD-10 Q20-Q26, and with corresponding ICD-8 and ICD-9 codes³² (see Supplementary data online, Table S2). In line with EUROCAT, we did not consider minor defects such as patent ductus arteriosus in preterm babies (<37 weeks) and isolated patent foramen ovale as major CHDs.³³

Secondary outcomes were (i) severe CHDs, (ii) CHDs according to the hierarchical classification of Botto et al.,³⁴ and (iii) 10 selected specific major CHDs. The severe CHD subgroup consists of 26 major CHDs classified as severe CHDs according to EUROCAT, which is based on Dolk et al.³⁵ (see Supplementary data online, Table S2). The Botto classification of CHDs has been designed for use in aetiological and observational studies.^34,36,37 Here, all CHDs are grouped in a hierarchical arrangement into six lesion groups, lesion group 1 being the most severe (see Supplementary data online, Table S3). Lesion group 1 includes conotruncal defects (such as tetralogy of Fallot, transposition of the great vessels, common arterial trunk, and aortopulmonary septal defects); lesion group 2 includes non-conotruncal defects [such as endocardial cushion defects, a common ventricle, and hypoplastic left heart syndrome (HLHS)]; lesion group 3 coarctation of the aorta; lesion group 4 ventricular septal defects (VSDs); and lesion group 5 atrial septal defects. Lesion group 6 includes all other CHD diagnoses and circulatory system anomalies not included in lesion groups 1–5. In individuals with several CHDs, only the most severe CHD was included. The 10 selected specific major CHDs were common arterial truncus, double outlet right ventricle, complete transposition of the great vessel, isomerism of atrial appendages with asplenia or polysplenia, atrioventricular septal defect, tetralogy of Fallot, pulmonary valve atresia, tricuspid atresia and stenosis, HLHS, and coarctation aortae (see Supplementary data online, Table S2). These specific CHDs were mainly selected based on medical knowledge and previous studies.^38,39

This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.⁴⁰

Statistical methods

Logistic regression analysis was performed estimating crude and adjusted ORs for CHD, with 95% CIs. We estimated the risk of major CHDs, severe CHDs, the 6 CHD lesion groups, and the 10 selected major CHDs within 1 year of age for Danish, Finnish, and Swedish children, and by the time of birth for Norwegian children since data on CHDs in Norway were available solely from the birth records.

Children born after ART were compared with children conceived spontaneously. Multiples included twins, triplets, and higher-order multiple births. Moreover, ART-conceived singletons were compared with spontaneously conceived singletons, ART-conceived multiples with spontaneously conceived multiples, ART-conceived multiples with ART-conceived singletons, and spontaneously conceived multiples with spontaneously conceived singletons. Where data were available (Denmark, Norway, and Sweden), we further compared singletons conceived using ICSI with singletons conceived using conventional IVF, and singletons conceived using frozen embryos with singletons conceived using fresh embryos. Singletons born after ICSI, IVF, or frozen embryo transfer (FET) were also compared with spontaneously conceived singletons.

The choice of covariates was based on previous studies and a thorough consideration of the existing knowledge of risk factors.^38,41–44 Adjustments were made for child’s year of birth (continuous variable), country of birth, maternal age at delivery (continuous variable), parity (nulliparous/parous), maternal smoking, maternal pre-gestational diabetes types 1 and 2, and history of maternal non-chromosomal CHDs. There were few missing data for most of these covariates except for smoking (Table 1). Missing data for smoking were set as no smoking in the main analysis. No imputation was conducted for other missing data. The significance level was set to 5%. All data analyses were calculated using STATA (version 18).

Table 1

Characteristics of study population by mode of conception and country of birth (Denmark 1994–2014, Finland 1990–2014, Norway 1984–2015, and Sweden 1987–2015)

	All countries		Denmark		Finland		Norway		Sweden
	n = 7 747 637 liveborn children		n = 1 355 267 liveborn children		n = 1 497 341 liveborn children		n = 1 865 484 liveborn children		n = 3 029 545 liveborn children
	ART	SC	ART	SC	ART	SC	ART	SC	ART	SC
	n = 171 735	n = 7 575 902	n = 45 801	n = 1 309 466	n = 29 691	n = 1 467 650	n = 34 042	n = 1 831 442	n = 62 201	n = 2 967 344
Child characteristics
Year of birth, n (%)
1984–90	1610 (0.9)	898 290 (11.9)	0	0	53 (0.2)	65 393 (4.5)	877 (2.6)	383 757 (21.0)	680 (1.1)	449 140 (15.1)
1991–95	11 681 (6.8)	1 321 614 (1.4)	1299 (2.8)	138 138 (10.6)	2864 (9.7)	321 476 (21.9)	2464 (7.2)	297 356 (16.2)	5054 (8.1)	564 644 (19.0)
1996–2000	28 709 (16.7)	1 333 763 (17.6)	8532 (18.6)	327 305 (25.0)	6941 (23.4)	283 333 (19.3)	4313 (12.7)	292 015 (15.9)	8923 (14.4)	431 110 (14.5)
2001–05	36 092 (21.0)	1 332 414 (17.6)	11 798 (25.8)	313 094 (23.9)	6359 (21.4)	276 400 (18.8)	6586 (19.4)	276 926 (15.1)	11 349 (18.3)	465 994 (15.7)
2006–10	45 519 (26.5)	1 414 225 (18.7)	13 156 (28.7)	310 143 (23.7)	6974 (23.5)	291 949 (19.9)	9411 (27.7)	293 014 (16.0)	15 978 (25.7)	519 119 (17.5)
2011–15	48 124 (28.0)	1 275 596 (16.8)	11 016 (24.1)	220 786 (16.9)	6500 (21.9)	229 099 (15.6)	10 391 (30.5)	288 374 (15.8)	20 217 (32.5)	537 337 (18.1)
Birth weight, n (%)
Very low, <1500 g	5215 (3.1)	54 992 (0.7)	1566 (3.5)	10 331 (0.8)	845 (2.9)	9504 (0.7)	1233 (3.6)	14 584 (0.8)	1571 (2.5)	20 573 (0.7)
Low, <2500, g	27448 (16.0)	319 685 (4.2)	8463 (18.5)	61 229 (4.7)	4783 (16.1)	56 334 (3.8)	6103 (17.9)	81 588 (4.5)	8099 (13.0)	120 534 (4.1)
Macrosomia, ≥4000g	20 516 (12.0)	1 425 820 (18.8)	4725 (10.3)	238 103 (18.2)	3521 (11.9)	272 464 (18.6)	3933 (11.6)	355 748 (19.4)	8337 (13.4)	559 505 (18.9)
Missing data on birth weight	755 (0.4)	36 271 (0.5)	463 (1.0)	24 570 (1.9)	13 (0.04)	3405 (0.2)	35 (0.1)	1556 (0.1)	244 (0.4)	6740 (0.2)
Gestational age, n (%)
Extremely preterm, <28 + 0 weeks	1984 (1.2)	20 482 (0.3)	626 (1.4)	3640 (0.3)	325 (1.1)	3613 (0.3)	455 (1.3)	5352 (0.3)	578 (0.9)	7877 (0.3)
Very preterm, <32 + 0 weeks	6120 (3.6)	64 821 (0.9)	1869 (4.1)	12 178 (0.9)	970 (3.3)	10 834 (0.7)	1400 (4.1)	16 984 (0.9)	1881 (3.0)	24 825 (0.8)
Preterm, <37 + 0 weeks	31 367 (18.3)	436 471 (5.8)	9264 (20.2)	78 548 (6.0)	5725 (19.3)	77 520 (5.3)	6867 (20.2)	108 524 (5.9)	9511 (15.3)	171 879 (5.8)
Post-term, ≥42 + 0 weeks	3296 (5.3)	513 869 (6.8)	1491 (3.3)	82 959 (6.3)	924 (3.1)	67 415 (4.6)	1557 (4.6)	158 092 (8.6)	3296 (5.3)	205 403 (6.9)
Missing data on gestational age	652 (0.4)	128 984 (1.7)	298 (0.7)	28 949 (2.2)	57 (0.2)	6755 (0.5)	246 (0.7)	89 599 (4.9)	51 (0.1)	3681 (0.1)
Plurality, n (%)
Singletons	127 275 (74.1)	7 380 916 (97.4)	31 077 (67.9)	1 269 638 (97.0)	22 044 (74.2)	1 431 723 (97.6)	24 117 (70.8)	1 783 926 (97.4)	50 037 (80.4)	2 895 629 (97.6)
Twins	42 459 (24.7)	190 426 (2.5)	14 332 (31.3)	38 810 (3.0)	7198 (24.2)	35 222 (2.4)	9364 (27.5)	46 356 (2.5)	11 565 (18.6)	70 038 (2.4)
Triplets and higher	2001 (1.2)	4504 (0.1)	392 (0.9)	1018 (0.1)	449 (1.5)	705 (0.1)	561 (1.7)	1160 (0.1)	599 (1.0)	1621 (0.1)
Missing data on plurality, n	0	56	0	0	0	0	0	0	0	56
Male sex, n (%)	87 785 (51.1)	3 887 791 (51.3)	23 211 (50.7)	672 079 (51.3)	15 197 (51.2)	749 952 (51.1)	17 456 (51.3)	940 571 (51.4)	31 921 (51.3)	1 525 189 (51.4)
Missing data on sex, n	0	7	0	0	0	0	0	0	0	7
Maternal characteristics
Age at delivery years, mean (SD)	33.4 (4.2)	29.3 (5.2)	33.1 (4.2)	29.7 (4.9)	33.4 (4.7)	29.3 (5.3)	33.2 (4.7)	28.8 (5.2)	33.7 (4.2)	29.3 (5.2)
Age at delivery (years), n (%)
<25	2773 (1.6)	1 401 083 (18.5)	715 (1.6)	186 390 (14.2)	695 (2.3)	280 446 (19.1)	506 (1.5)	394 183 (21.5)	857 (1.4)	540 064 (18.2)
25–29	28 888 (16.8)	2 574 096 (34.0)	8278 (18.1)	454 202 (34.7)	5424 (18.3)	489 753 (33.4)	6098 (17.9)	637 419 (34.8)	9088 (14.6)	992 722 (33.5)
30–34	70 849 (41.3)	2 371 342 (31.3)	19 848 (43.3)	453 307 (34.6)	11 776 (39.7)	446 451 (30.4)	14 402 (42.3)	538 859 (29.4)	24 823 (39.9)	932 725 (31.4)
35–39	56 593 (33.0)	1 027 140 (13.6)	13 813 (30.2)	184 443 (14.1)	8956 (30.2)	204 704 (14.0)	11 203 (32.9)	221 121 (12.1)	22 621 (36.4)	416 872 (14.1)
>40	12 632 (7.4)	202 241 (2.7)	3147 (6.9)	31 124 (2.4)	2840 (9.6)	46 296 (3.2)	1833 (5.4)	39 860 (2.2)	4812 (7.7)	84 961 (2.9)
Primiparous, n (%)	116 520 (67.9)	3 163 682 (41.8)	30 490 (66.6)	554 283 (42.3)	2138 (67.8)	593 361 (40.4)	21 733 (63.8)	756 578 (41.3)	44 159 (71.0)	1 259 460 (42.4)
Missing data on parity, n (%)	569 (0.33)	19 165 (0.3)	560 (1.2)	17 719 (1.4)	9 (0.03)	1374 (0.1)	0	0	0	0
BMI (kg/m2), n (% of non-missing)
<18.5	2565 (2.4)	132 058 (3.5)	942 (3.4)	25 980 (4.2)	391 (2.6)	22 448 (3.7)	310 (0.2)	10 619 (4.1)	922 (1.7)	73 011 (3.1)
18.5–24.9	68 169 (63.4)	2 415 098 (63.4)	18 050 (64.6)	385 134 (62.5)	9840 (65.3)	375 550 (62.6)	5878 (61.0)	158 114 (61.7)	34 411 (62.8)	1 496 300 (64.0)
25.0–29.9	25 989 (24.2)	851 663 (22.4)	6196 (22.2)	129 398 (21.0)	3302 (21.9)	130 144 (21.7)	2298 (23.8)	56 411 (22.0)	14 193 (25.9)	535 710 (22.9)
≥30	10 746 (10.0)	410 734 (10.8)	2774 (9.9)	75 920 (12.3)	1543 (10.2)	72 271 (12.0)	1154 (12.0)	31 139 (12.2)	5275 (9.6)	231 404 (9.9)
Missing data for BMI, n (% of total)	64 256 (37.4)	3 766 349 (49.7)	17 839 (39.0)	693 037 (52.9)	14 615 (49.2)	867 237 (59.1)	24 402 (71.7)	1 575 159 (86.0)	7400 (11.9)	630 919 (21.3)
Smoking, n (%)	10 120 (5.9)	934 433 (12.3)	3956 (8.6)	194 514 (14.9)	1815 (6.1)	224 950 (15.3)	1735 (5.1)	129 171 (7.1)	2614 (4.2)	385 798 (13.0)
Missing data on smoking, n (%)	12 938 (7.5)	1 199 059 (15.8)	3217 (7.0)	132 470 (10.1)	449 (1.5)	37 317 (2.5)	5703 (16.8)	868 891 (47.4)	3569 (5.7)	160 381 (5.4)
Educational levela,b, n (%)
Low	59 160 (43.0)	2 972 924 (51.7)	24 096 (52.6)	775 820 (59.3)	8675 (29.2)	604 510 (41.2)	NA	NA	26 389 (42.4)	1 592 594 (53.7)
Middle	45 314 (32.9)	1 604 375 (27.9)	13 662 (29.8)	342 853 (26.2)	11 349 (38.2)	466 541 (31.8)	NA	NA	20 303 (32.6)	794 981 (26.8)
High	27 952 (20.3)	765 000 (13.3)	7 463 (16.3)	154 271 (11.8)	8078 (27.2)	237 162 (16.2)	NA	NA	12 411 (20.0)	373 567 (12.6)
Missing data on education, n (%)	5267 (3.8)	402 161 (7.0)	580 (1.3)	36 522 (2.8)	1589 (5.4)	159 437 (10.9)	NA	NA	3098 (5.0)	206 202 (6.7)
Pre-gestational diabetes n (%)	1515 (0.88)	53 709 (0.71)	517 (1.13)	16 105 (1.23)	372 (1.25)	14 266 (0.97)	268 (0.79)	9352 (0.51)	358 (0.58)	13 986 (0.47)
Non-chromosomal CHD, n (%)	648 (0.38)	26 017 (0.34)	167 (0.36)	4649 (0.36)	159 (0.54)	7868 (0.54)	54 (0.16)	1739 (0.09)	268 (0.43)	11 761 (0.40)
Chromosomal defects, n (%)	661 (0.38)	5925 (0.08)	232 (0.51)	1368 (0.10)	107 (0.36)	2066 (0.14)	28 (0.08)	872 (0.05)	294 (0.47)	1619 (0.05)
Paternal characteristics
Non-chromosomal CHD, n (%)	NA	NA	NA	NA	NA	NA	NA	NA	243 (0.14)	10 448 (0.35)
Chromosomal defects, n (%)	NA	NA	NA	NA	NA	NA	NA	NA	217 (0.35)	768 (0.03)
Assisted reproduction methodc, n (%)
IVF	81 900 (57.7)	-	25 024 (54.6)	-	NA	-	20 093 (59.0)	-	36 783 (59.1)	-
ICSI	55 126 (38.8)	-	18 118 (39.6)	-	NA	-	11 590 (34.1)	-	25 418 (40.9)	-
Missing data on IVF/ICSI	5018 (3.5)	-	2659 (5.8)	-	NA	-	2359 (6.9)	-	0	-
Single embryo transfer	52 130 (36.7)	-	10 191 (22.3)	-	NA	-	10 303 (30.3)	-	31 636 (50.9)
Fresh embryo transfer	115 408 (81.2)	-	41 022 (89.6)	-	NA	-	25 630 (75.3)	-	48 756 (78.4)	-
Frozen embryo transfer	22 630 (15.9)	-	4761 (10.4)	-	NA	-	4424 (13.0)	-	13 445 (21.6)	-
Missing data on fresh/frozen cycle	4006 (2.8)	-	18 (0.04)	-	NA	-	3988 (11.7)	-	0	-
Cleavage stage embryo	130 079 (91.6)	-	43 621 (95.2)	-	NA	-	30 219 (88.8)	-	56 239 (90.4)	-
Blastocysts	7663 (5.4)	-	1287 (2.8)	-	NA	-	414 (1.2)	-	5962 (9.6)	-
Missing data on embryo stage	4302 (3.0)	-	893 (2.0)	-	NA	-	3409 (10.0)	-	0	-
Donated oocytes	1410 (1.0)	-	710 (1.6)	-	NA	-	0e	-	700 (1.1)	-
Donated spermd	2994 (2.8)	-	1911 (4.2)	-	NA	-	NA	-	1083 (1.7)	-

ART, assisted reproductive technology; BMI, body mass index; CHD, congenital heart defect; ICSI, intracytoplasmic sperm injection; IVF, in vitro fertilization; NA, not available; SC, spontaneous conception.

^aData for Denmark, Finland, and Sweden (denominator ART, 137 693; SC, 5 744 460).

^bHighest educational level according to the International Standard Classification of Education (ISCED2011): ISCED < 5, primary, secondary, or post-secondary non tertiary education; ISCED 5–6, first stage of tertiary education (bachelor or equivalent); and ISCED 7–8, second stage of tertiary education (master, doctorate, or more).⁴⁵.

^cData for Denmark, Norway, and Sweden (denominator ART = 142 044).

^dData for Denmark and Sweden (denominator ART = 108 002).

^eOocyte donation not permitted in Norway during the study period.

Sensitivity analyses

We performed several sensitivity analyses on major CHDs. First, we performed a sensitivity analysis restricting the analysis to those with known data on smoking. Second, maternal highest educational level (low, medium, and high)⁴⁵ was included as a covariate when this information was available (Denmark, Finland, and Sweden). Third, a sensitivity analysis was performed including Finnish data with validated data on major CHDs and excluding those which after further evaluation by the Finnish malformation registry were considered as minor CHDs (see Supplementary data online, Table S1).⁴⁶ A fourth sensitivity analysis on major CHDs and severe CHDs was performed restricting the cohort to infants born 2006–15. Between 2004 and 2007, all countries had introduced the second trimester ultrasound including foetal anomaly screening (gestational week 18–21). Lastly, a sensitivity analysis was performed on Swedish data, adjusting also for paternal CHD.

Results

Baseline characteristics

A total of 7 747 637 liveborn children were included, 171 735 (2.2%) were born after ART and 7 575 902 (97.8%) were born after SC (Figure 1). The proportion of multiples was 25.9% (44 460 of 171 735) in the ART group and 2.6% (194 930 of 7 575 902) in the spontaneously conceived group.

Figure 1

Flow chart of the study population

Baseline characteristics are presented in Table 1. Children conceived by ART were more often born preterm (<37 weeks; 18.3% vs. 5.8%) or with LBW (<2500 g; 16.0% vs. 4.2%) than children born after SC. Women who conceived by ART were more likely to be older at birth (≥35 years; 40.4% vs. 16.3%), to be primiparous (67.9% vs. 41.8%), to have a high educational level (16.3% vs. 10.1%), and to be non-smokers (86.6% vs. 71.9%). The prevalence of pre-gestational diabetes was 0.88% in women who conceived using ART vs. 0.71% in women who conceived spontaneously. Corresponding figures for maternal non-chromosomal CHDs were 0.38% vs. 0.34% and for obesity (body mass index ≥ 30 kg/m²) 10.0% vs. 10.8%.

Major congenital heart defects

The rate of major CHDs varied somewhat between countries as indicated in Table 2 and Figure 2.

Figure 2

Forest plot showing the ORs for independent covariates of risk of major congenital heart defects in children born after assisted reproductive technology vs. children conceived spontaneously. Footnote: Please observe that in this figure, child’s year of birth and maternal age are expressed in categories for better visualization. In all statistical models, however, both variables are analysed as continuous variables, resulting in slightly different AOR and 95% CI for comparison ART vs. SC

Table 2

Risk of congenital heart defects in liveborn children conceived by assisted reproductive technology vs. spontaneous conception for all countries and by country (Denmark 1994–2014, Finland 1990–2014, Norway 1984–2015, and Sweden 1987–2015)

	No. of liveborn children		Risk of CHD, ART vs. spontaneous conception
	ART	Spontaneous conception	Crude OR (95% CI) P value	Adjusted ORa (95% CI) P value
Major CHDsb diagnosed within the first year of life
All countries n/N (%)	3159/171 735 (1.84)	86 824/7 575 902 (1.15)	1.62 (1.56–1.68) < .001	1.36 (1.31–1.41) < .001
Denmark n/N (%)	879/45 801 (1.92)	14 467/1 309 466 (1.10)	1.75 (1.64–1.88) < .001	1.70 (1.58–1.82) < .001
Finland n/N (%)	592/29 691 (1.99)	19 341/1 467 650 (1.32)	1.52 (1.40–1.65) < .001	1.33 (1.22–1.45) < .001
Norway (at birth) n/N (%)	380/34 042 (1.12)	10 630/1 831 442 (0.58)	1.93 (1.74–2.14) < .001	1.34 (1.21–1.49) < .001
Sweden n/N (%)	1308/62 201 (2.10)	42 386/2 967 344 (1.43)	1.48 (1.40–1.57) < .001	1.22 (1.16–1.29) < .001
Severe CHDsb diagnosed within the first year of life
All countries n/N (%)	594/171 735 (0.35)	19 375/7 575 902 (0.26)	1.35 (1.25–1.47) < .001	1.30 (1.20–1.42) < .001
Denmark n/N (%)	137/45 801 (0.30)	3042/1 309 466 (0.23)	1.29 (1.09–1.53) .004	1.28 (1.07–1.53) .006
Finland n/N (%)	117/29 691 (0.39)	4320/1 467 650 (0.29)	1.34 (1.11–1.61) .002	1.27 (1.06–1.54) .011
Norway (at birth) n/N (%)	90/34 042 (0.26)	3642/1 831 442 (0.20)	1.33 (1.08–1.64) .008	1.23 (0.99–1.52) .060
Sweden n/N (%)	250/62 201 (0.40)	8371/2 967 344 (0.28)	1.43 (1.26–1.62) < .001	1.37 (1.20–1.56) < .001
Major CHDs diagnosed within the first year of life according to the hierarchic classification of Botto et al.c
Lesion group 1 conotruncal
All countries n/N (%)	194/171 735 (0.11)	6314/7 575 902 (0.08)	1.36 (1.18–1.56) < .001	1.23 (1.06–1.42) .006
Denmark n/N (%)	49/45 801 (0.11)	1255/1 309 466 (0.10)	1.12 (0.84–1.49) .450	1.14 (0.86–1.52) .366
Finland n/N (%)	36/29 691 (0.12)	1205/1 467 650 (0.08)	1.48 (1.06–2.06) .021	1.46 (1.04–2.04) .027
Norway (at birth) n/N (%)	35/34 042 (0.10)	965/1 831 442 (0.05)	1.95 (1.39–2.74) < .001	1.45 (1.03–2.04) .033
Sweden n/N (%)	74/62 201 (0.12)	2889/2 967 344 (0.10)	1.22 (0.97–1.54) .088	1.20 (0.95–1.51) .134
Lesion group 2 Non-conotruncal
All countries n/N (%)	158/171 735 (0.09)	5700/7 575 902 (0.08)	1.22 (1.04–1.43) .013	1.27 (1.08–1.50) .003
Denmark n/N (%)	42/45 801 (0.09)	835/1 309 466 (0.06)	1.44 (1.05–1.96) .022	1.40 (1.02–1.92) .035
Finland n/N (%)	34/29 691 (0.11)	1038/1 467 650 (0.07)	1.62 (1.15–2.28) .006	1.45 (1.02–2.04) .036
Norway (at birth) n/N (%)	32/34 042 (0.09)	1787/1 831 442 (0.10)	0.96 (0.68–1.37) .834	1.19 (0.84–1.70) .325
Sweden n/N (%)	50/62 201 (0.08)	2040/2 967 344 (0.07)	1.17 (0.88–1.55) .274	1.12 (0.84–1.48) .441
Lesion group 3 coarctation aortae
All countries n/N (%)	105/171 735 (0.06)	3397/7 575 902 (0.04)	1.36 (1.12–1.66) .002	1.22 (0.999–1.49) .051
Denmark n/N (%)	20/45 801 (0.04)	522/1 309 466 (0.04)	1.10 (0.70–1.71) .689	1.06 (0.67–1.66) .815
Finland n/N (%)	19/29 691 (0.06)	1081/1 467 650 (0.07)	0.87 (0.55–1.37) .543	0.79 (0.50–1.24) .305
Norway (at birth) n/N (%)	10/34 042 (0.03)	313/1 831 442 (0.02)	1.72 (0.92–3.23) .092	1.14 (0.60–2.15) .689
Sweden n/N (%)	56/62 201 (0.09)	1481/2 967 344 (0.05)	1.80 (1.38–2.36) < .001	1.57 (1.20–2.06) .001
Lesion group 4 VSD
All countries n/N (%)	1212/171 735 (0.71)	36 219/7 575 902 (0.48)	1.48 (1.40–1.57) < .001	1.21 (1.14–1.29) < .001
Denmark n/N (%)	185/45 801 (0.40)	3812/1 309 466 (0.29)	1.39 (1.20–1.61) < .001	1.39 (1.20–1.62) < .001
Finland n/N (%)	333/29 691 (1.12)	10 677/1 467 650 (0.73)	1.55 (1.39–1.73) < .001	1.39 (1.25–1.56) < .001
Norway (at birth)	153/34 042 (0.45)	3661/1 831 442 (0.20)	2.25 (1.92–2.65) < .001	1.41 (1.20–1.66) < .001
Sweden n/N (%)	541/62 201 (0.87)	18 069/2 967 344 (0.61)	1.43 (1.31–1.56) < .001	1.16 (1.06–1.26) .001
Lesion group 5 ASD
All countries n/N (%)	720/171 735 (0.42)	13 787/7 575 902 (0.18)	2.31 (2.14–2.49) < .001	1.60 (1.48–1.73) < .001
Denmark n/N (%)	288/45 801 (0.63)	3372/1 309 466 (0.26)	2.45 (2.17–2.77) < .001	2.21 (1.96–2.50) < .001
Finland n/N (%)	23/29 691 (0.08)	1038/1 467 650 (0.07)	1.10 (0.72–1.66) .666	0.90 (0.59–1.37) .623
Norway (at birth) n/N (%)	82/34 042 (0.24)	1559/1 831 442 (0.09)	2.83 (2.27–3.54) < .001	1.75 (1.40–2.19) < .001
Sweden n/N (%)	327/62 201 (0.53)	7818/2 967 344 (0.26)	2.00 (1.79–2.24) < .001	1.38 (1.23–1.54) < .001
Lesion group 6 other CHDs
All countries n/N (%)	770/171 735 (0.45)	21 407/7 575 902 (0.28)	1.59 (1.48–1.71) < .001	1.48 (1.37–1.59) < .001
Denmark n/N (%)	295/45 801 (0.64)	4671/1 309 466 (0.36)	1.81 (1.61–2.04) < .001	2.09 (1.85–2.35) < .001
Finland n/N (%)	147/29 691 (0.50)	4302/1 467 650 (0.29)	1.69 (1.44–2.00) < .001	1.50 (1.27–1.77) < .001
Norway (at birth) n/N (%)	68/34 042 (0.20)	2345/1 831 442 (0.13)	1.56 (1.23–1.99) < .001	1.13 (0.88–1.44) .332
Sweden n/N (%)	260/62 201 (0.42)	10 089/2 967 344 (0.34)	1.23 (1.09–1.39) .001	1.27 (1.12–1.44) < .001

ART, assisted reproductive technology; ASD, atrial septal defect; CHD, congenital heart defects; CI, confidence interval; OR, odds ratio; VSD, ventricular septal defect.

^aMajor CHDs and severe CHDs: Adjustment for child’s year of birth, country of birth, maternal age, parity, maternal smoking, maternal diabetes, and maternal CHD, in the analysis of all countries. Adjustment for child’s year of birth, maternal age, parity, maternal smoking, maternal diabetes, and maternal CHD, in the analysis of the specific countries. Botto lesion groups 1–6: Adjustment for child’s year of birth, country of birth, maternal age, parity, maternal smoking, maternal diabetes, and maternal CHD in the analysis of all countries. Only adjustment for child’s year of birth and maternal age in the analysis of the specific countries.

^bMajor CHDs and severe CHDs according to the EUROCAT 1.5 definition.^32,33.

^cLesion groups 1–6 according to Botto et al.³⁴

Major CHDs diagnosed up to 1 year of age were detected in 3159 children born after ART (1.84%) and in 86 824 children born after SC (1.15%; adjusted OR 1.36; 95% CI 1.31–1.41; P < .001; Table 2). Among children with major CHDs, 193 children (6.1%) in the ART group and 5472 (6.3%) in the spontaneously conceived group had a concomitant chromosomal aberration. Associations of each covariate with major CHDs are illustrated in Figure 2. The strongest associations were seen for maternal pre-gestational diabetes (OR 2.72; 95% CI 2.59–2.86) and maternal CHDs (OR 3.80; 95% CI 3.59–4.02).

Major CHDs were detected among 1.62% (n = 2059) of singletons born after ART and among 1.11% (n = 82 119) of singletons born after SC (adjusted OR 1.19; 95% CI 1.14–1.24; P < .001; Table 3). No significant difference was seen for multiples conceived after ART vs. multiples conceived after SC (Table 3).

Table 3

Risk of congenital heart defects in singletons conceived by assisted reproductive technology vs. spontaneous conception and multiples conceived by assisted reproductive technology vs. spontaneous conception (Denmark 1994–2014, Finland 1990–2014, Norway 1984–2015, and Sweden 1987–2015)

	No. of singletons		No. of multiples		Risk of CHD in ART singletons vs. singletons born after spontaneous conception		Risk of CHD in ART multiples vs. multiples born after spontaneous conception
	ART n = 127 275	Spontaneous conception n = 7 380 916	ART n = 44 460	Spontaneous conception n = 194 930	Crude OR (95% CI) P value	Adjusted ORa (95% CI) P value	Crude OR (95% CI) P value	Adjusted ORa (95% CI) P value
Major CHDsb diagnosed within the first year of life, n (%)	2059 (1.62)	82 119 (1.11)	1100 (2.47)	4707 (2.41)	1.46 (1.40–1.53) < .001	1.19 (1.14–1.24) < .001	1.03 (0.96–1.10) .455	0.94 (0.88–1.01) .085
Severe CHDsb diagnosed within the first year of life, n (%)	399 (0.31)	18 539 (0.25)	195 (0.44)	836 (0.43)	1.25 (1.13–1.38) < .001	1.20 (1.09–1.33) < .001	1.02 (0.87–1.20) .778	0.97 (0.82–1.14) .669
Major CHDs diagnosed within the first year of life according to the hierarchic classification of Botto et al.c
Lesion group 1 conotruncal, n (%)	138 (0.11)	6057 (0.08)	56 (0.13)	257 (0.13)	1.32 (1.12–1.56) .001	1.20 (1.01–1.42) .042	0.96 (0.72–1.28) .757	0.84 (0.62–1.24) .273
Lesion group 2 non-conotruncal, n (%)	99 (0.08)	5499 (0.07)	59 (0.13)	201 (0.10)	1.04 (0.86–1.27) .671	1.12 (0.92–1.37) .260	1.29 (0.96–1.72) .088	1.14 (0.83–1.55) .420
Lesion group 3 coarctation aortae, n (%)	64 (0.05)	3221 (0.04)	41 (0.09)	176 (0.09)	1.15 (0.90–1.48) .261	1.00 (0.78–1.28) .988	1.02 (0.73–1.44) .903	1.03 (0.72–1.49) .855
Lesion group 4 VSD, n (%)	882 (0.69)	34 552 (0.47)	330 (0.74)	1667 (0.86)	1.48 (1.39–1.59) < .001	1.15 (1.07–1.23) < .001	0.87 (0.77–0.98) .018	0.90 (0.79–1.02) .086
Lesion group 5 ASD, n (%)	427 (0.34)	12 569 (0.17)	293 (0.66)	1218 (0.62)	1.97 (1.79–2.17) < .001	1.30 (1.17–1.43) < .001	1.06 (0.93–1.20) .412	0.95 (0.83–1.09) .400
Lesion group 6 other CHDs, n (%)	449 (0.35)	20 221 (0.27)	321 (0.72)	1186 (0.61)	1.29 (1.17–1.42) < .001	1.20 (1.09–1.32) < .001	1.19 (1.05–1.34) .006	1.01 (0.88–1.15) .879

ART, assisted reproductive technology; ASD, atrial septal defect; CHD, congenital heart defect; CI, confidence interval; OR, odds ratio; VSD, ventricular septal defect.

^aAdjustment for child’s year of birth, country of birth, maternal age, parity, maternal smoking, maternal diabetes, and maternal CHD.

^bMajor CHDs and severe CHDs according to the EUROCAT 1.5 definition.^32,33.

^cLesion groups 1–6 according to Botto et al.³⁴

Multiples born after ART had an absolute risk of major CHDs of 2.47% (n = 1100; adjusted OR 1.70; 95% CI 1.58–1.84; P < .001 vs. singletons conceived after ART; Table 4). Multiples born after SC had an absolute risk of major CHDs of 2.41% (n = 4705; adjusted OR 2.17; 95% CI 2.10–2.24; P < .001 vs. spontaneously conceived singletons; Table 4).

Table 4

Risk of congenital heart defects in multiples born after ART vs. singletons born after assisted reproductive technology and multiples born after spontaneous conception vs. singletons born after spontaneous conception (Denmark 1994–2014, Finland 1990–2014, Norway 1984–2015, and Sweden 1987–2015)

	No of singletons		No of multiples		Risk of CHD in ART multiples vs. ART singletons		Risk of CHD in SC multiples vs. SC singletons
	ART n = 127 275	SC n = 7 380 916	ART n = 44 460	SC n =194 930	Crude OR (95% CI) P value	Adjusted OR* (95% CI) P value	Crude OR (95% CI) P value	Adjusted ORa (95% CI) P value
Major CHDsb diagnosed within the first year of life, n (%)	2059 (1.62)	82 119 (1.11)	1100 (2.47)	4705 (2.41)	1.54 (1.43–1.66) < .001	1.70 (1.58–1.84) < .001	2.20 (2.13–2.26) < .001	2.17 (2.10–2.24) < .001
Severe CHDb diagnosed within the first year of life, n (%)	399 (.31)	18 539 (0.25)	195 (0.44)	836 (0.43)	1.40 (1.18–1.66) < .001	1.46 (1.22–1.75) < .001	1.71 (1.60–1.83) < .001	1.70 (1.58–1.82) < .001
Major CHDs diagnosed within the first year of life according to the hierarchic classification of Botto et al.c
Lesion group 1 conotruncal, n (%)	138 (0.11)	6057 (0.08)	56 (0.13)	257 (0.13)	1.16 (0.85–1.59) .344	1.19 (0.86–1.65) .281	1.61 (1.42–1.82) < .001	1.58 (1.39–1.79) < .001
Lesion group 2 non-conotruncal, n (%)	99 (0.08)	5499 (0.07)	59 (0.13)	201 (0.10)	1.71 (1.24–2.36) .001	1.63 (1.17–2.28) .004	1.38 (1.20–1.59) < .001	1.37 (1.19–1.57) < .001
Lesion group 3 coarctation aortae, n (%)	64 (0.05)	3221 (0.04)	41 (0.09)	176 (0.09)	1.83 (1.24–2.72) .002	2.05 (1.36–3.11) .001	2.07 (1.78–2.41) < .001	2.05 (1.76–2.39) < .001
Lesion group 4 VSD, n (%)	882 (0.69)	34 552 (0.47)	330 (0.74)	1667 (0.86)	1.07 (0.94–1.22) .286	1.26 (1.10–1.44) .001	1.83 (1.75–1.93) < .001	1.83 (1.74–1.92) < .001
Lesion group 5 ASD, n (%)	427 (0.34)	12 569 (0.17)	293 (0.66)	1218 (0.62)	1.97 (1.70–2.29) < .001	2.49 (2.13–2.92) < .001	3.69 (3.47–3.91) < .001	3.50 (3.30–3.71) < .001
Lesion group 6 Other CHDs, n (%)	449 (0.35)	20 221 (0.27)	321 (0.72)	1186 (0.61)	2.05 (1.78–2.37) < .001	1.93 (1.66–2.34) < .001	2.23 (2.10–2.36) < .001	2.21 (2.08–2.34) < .001

ART, assisted reproductive technology; ASD, atrial septal defect; CHD, congenital heart defect; CI, confidence interval; OR, odds ratio; SC, spontaneous conception; VSD, ventricular septal defect.

^aAdjustment for child’s year of birth, country of birth, maternal age, parity, maternal smoking, maternal diabetes, and maternal CHD.

^bMajor CHDs and severe CHDs according to the EUROCAT 1.5 definition.^32,33.

^cLesion groups 1–6 according to Botto et al.³⁴

Table 5 shows the results for singletons born after ICSI (n = 42 385), IVF (n = 59 244), and SC (n = 5 949 193). Major CHDs were detected among 1.67% (n = 709) of singletons born after ICSI and among 1.51% (n = 895) singletons born after IVF (adjusted OR 1.07; 95% CI 0.97–1.18; P = .200; Table 5).

Table 5

Risk of congenital heart defects by type of in vitro fertilization treatment (intracytoplasmic sperm injection or in vitro fertilization) in singletons conceived by assisted reproductive technology and spontaneous conception (Denmark 1994–2014, Norway 1984–2015, and Sweden 1987–2015)

	No of singletons			Risk of CHD, ICSI vs. IVF		Risk of CHD, ICSI vs. SC		Risk of CHD, IVF vs. SC
	ICSI n = 42 385	IVF n = 59 244	SC n = 5 949 193	Crude OR (95% CI) P value	Adjusteda OR (95% CI) P value	Crude OR (95% CI) P value	Adjustedb OR (95% CI) P value	Crude OR (95% CI) P value	Adjustedb OR (95% CI) P value
Major CHDsc  n (%)	709 (1.67)	895 (1.51)	63 675 (1.07)	1.11 (1.00–1.22) .041	1.07 (0.97–1.18) .200	1.57 (1.46–1.69) < .001	1.21 (1.12–1.31) < .001	1.42 (1.33–1.52) < .001	1.14 (1.07–1.22) < .001
Severe CHDsc  n (%)	124 (0.29)	180 (0.30)	14 390 (0.24)	0.96 (0.77–1.21) .746	0.93 (0.74–1.17) .536	1.21 (1.01–1.44) .035	1.17 (0.98–1.40) .085	1.26 (1.08–1.46) .002	1.19 (1.03–1.39) .019
Major CHDs diagnosed within the first year of life according to the hierarchic classification of Botto et al.d
Lesion group 1 conotruncal, n (%)	37 (0.09)	70 (0.12)	4896 (0.08)	0.74 (0.50–1.10) .136	0.73 (0.49–1.10) .130	1.06 (0.77–1.47) .721	0.94 (0.68–1.30) .688	1.44 (1.13–1.82) .003	1.28 (1.01–1.62) .046
Lesion group 2 non-conotruncal, n (%)	31 (0.07)	42 (0.07)	4500 (0.08)	1.03 (0.65–1.64) .895	1.03 (0.64–1.67) .889	0.97 (0.68–1.38) .852	1.14 (0.80–1.63) .459	0.94 (0.69–1.27) .676	1.00 (0.74–1.36) .995
Lesion group 3 coarctation aortae, n (%)	21 (0.05)	28 (0.05)	2185 (0.04)	1.05 (0.60–1.85) .870	0.99 (0.56–1.76) .977	1.35 (0.88–2.07) .172	1.06 (0.69–1.63) .795	1.29 (0.89–1.87) .185	1.07 (0.73–1.56) .733
Lesion group 4 VSD, n (%)	270 (0.64)	356 (0.60)	24 349 (0.41)	1.06 (0.90–1.24) .468	1.03 (0.87–1.21) .752	1.56 (1.38–1.76) < .001	1.08 (0.96–1.22) .198	1.47 (1.32–1.63) < .001	1.09 (0.98–1.21) .115
Lesion group 5 ASD, n (%)	191 (0.45)	209 (0.35)	11 598 (0.19)	1.28 (1.05–1.56) .014	1.18 (0.97–1.44) .101	2.32 (2.01–2.67) < .001	1.42 (1.23–1.64) < .001	1.81 (1.58–2.08) < .001	1.22 (1.06–1.41) .005
Lesion group 6 other CHDs, n (%)	159 (0.38)	190 (0.32)	16 147 (0.27)	1.17 (0.95–1.45) .144	1.17 (0.94–1.45) .163	1.38 (1.18–1.62) < .001	1.33 (1.14–1.56) < .001	1.18 (1.02–1.36) .022	1.14 (0.99–1.32) .079

ART, assisted reproductive technology; ASD, atrial septal defect; CHD, congenital heart defect; CI, confidence interval; ICSI, intracytoplasmic sperm injection; IVF, in vitro fertilisation; OR, odds ratio; SC, spontaneous conception; VSD, ventricular septal defect.

^aAdjustment for child’s year of birth, country of birth, maternal age, parity, maternal smoking, maternal diabetes, maternal CHD, and fresh and frozen embryo transfer.

^bAdjustment for child’s year of birth, country of birth, maternal age, parity, maternal smoking, maternal diabetes, and maternal CHD.

^cMajor CHDs and severe CHDs according to the EUROCAT 1.5 definition.^32,33.

^dLesion groups 1–6 according to Botto et al.³⁴

Table 6 presents the results for singletons born after frozen (n = 18 875) and fresh (n = 83 649) embryo transfer and SC (n = 5 949 193). The occurrence of major CHDs among singletons born after FET was 1.82% (n = 343) and among singletons born after fresh embryo transfer 1.54% (n = 1286; adjusted OR 1.04; 95% CI 0.91–1.18; P = .603).

Table 6

Risk of congenital heart defects by frozen and fresh embryo transfer in singletons conceived by assisted reproductive technology and spontaneous conception (Denmark 1994–2014, Norway 1984–2015, and Sweden 1987–2015)

	No of singletons			Risk of CHD, frozen vs. fresh embryo transfer		Risk of CHD, frozen embryo transfer vs. SC
	Frozen embryo transfer n = 18 875	Fresh embryo transfer n = 83 649	Spontaneous conception n = 5 949 193	Crude OR (95% CI) P value	Adjusteda OR (95% CI) P value	Crude OR (95% CI) P value	Adjustedb OR (95% CI) P value
Major CHDsc, n (%)	343 (1.82)	1286 (1.54)	63 675 (1.07)	1.19 (1.05–1.34) .006	1.04 (0.91–1.18) .603	1.71 (1.54–1.90) < .001	1.20 (1.08–1.34) .001
Severe CHDsc, n (%)	64 (0.34)	244 (0.29)	14 390 (0.24)	1.16 (0.88–1.53) .283	1.04 (0.77–1.41) .784	1.40 (1.10–1.79) .007	1.30 (1.02–1.67) .037
Major CHDs diagnosed within the first year of life according to the hierarchic classification of Botto et al.d
Lesion group 1 conotruncal, n (%)	24 (0.13)	85 (0.10)	4896 (0.08)	1.25 (0.80–1.97) .332	1.18 (0.72–1.95) .504	1.55 (1.03–2.31) .033	1.33 (0.89–1.98) .171
Lesion group 2 non-conotruncal, n (%)	15 (0.08)	58 (0.07)	4500 (0.08)	1.15 (0.65–2.02) .638	1.43 (0.77–2.66) .261	1.05 (0.63–1.74) .848	1.26 (0.76–2.09) .375
Lesion group 3 coarctatio aortae, n (%)	12 (0.06)	38 (0.05)	2185 (0.04)	1.40 (0.73–2.68) .310	1.13 (0.56–2.29) .726	1.73 (0.98–3.05) .058	1.22 (0.69–2.16) .488
Lesion group 4 VSD, n (%)	137 (0.73)	493 (0.59)	24 349 (0.41)	1.23 (1.02–1.49) .031	1.00 (0.82–1.23) .963	1.78 (1.50–2.11) < .001	1.06 (0.89–1.25) .516
Lesion group 5 ASD, n (%)	91 (0.48)	320 (0.38)	11 598 (0.19)	1.26 (1.00–1.59) .051	1.04 (0.80–1.34) .766	2.48 (2.02–3.05) < .001	1.33 (1.08–1.63) .008
Lesion group 6 other CHDs, n (%)	64 (0.34)	292 (0.35)	16 147 (0.27)	0.97 (0.74–1.27) .833	0.95 (0.70–1.29) .755	1.25 (0.98–1.60) .075	1.20 (0.93–1.53) .155

ART, assisted reproductive technology; ASD, atrial septal defect; CHD, congenital heart defect; CI, confidence interval; ICSI, intracytoplasmic sperm injection; IVF, in vitro fertilisation; OR, odds ratio; SC, spontaneous conception; VSD, ventricular septal defect.

^aAdjustment for child’s year of birth, country of birth, maternal age, parity, maternal smoking, maternal diabetes, maternal CHD, and IVF/ICSI.

^bAdjustment for child’s year of birth, country of birth, maternal age, parity, maternal smoking, maternal diabetes, and maternal CHD.

^cMajor CHDs and severe CHDs according to the EUROCAT 1.5 definition.^32,33.

^dLesion groups 1–6 according to Botto et al.³⁴

Severe congenital heart defects

Severe CHDs were detected in 594 children born after ART (0.35%) and in 19 375 children born after SC (0.26%; adjusted OR 1.30; 95% CI 1.20–1.42; P < .001; Table 2). Severe CHDs occurred among 0.31% (n = 399) singletons born after ART and among 0.25% (n = 18 539) singletons born after SC (adjusted OR 1.20; 95% CI 1.09–1.33; P < .001; Table 3). In multiples born after ART, the prevalence of severe CHDs was 0.44% (n = 195; adjusted OR 1.46; 95% CI 1.22–1.75; P < .001 vs. singletons conceived after ART; Table 4). In multiples born after SC, the prevalence of severe CHDs was 0.43% (n = 836; adjusted OR 1.70; 95% CI 1.58–1.82; P < .001 vs. spontaneously conceived singletons; Table 4). No significant difference in risk of severe CHDs was seen for multiples born after ART vs. multiples born after SC (Table 3).

Severe CHDs occurred among singletons born after ICSI in 0.29% (n = 124) and among singletons born after IVF in 0.30% (n = 180; adjusted OR 0.93; 95% CI 0.74–1.17; P = .536; Table 5). In singletons born after FET, the prevalence of severe CHDs was 0.34% (n = 64), and among singletons born after fresh embryo, transfer the prevalence was 0.29% (n = 244; adjusted OR 1.04; 95% CI 0.77–1.41; P = .784).

Sensitivity analyses

Information on smoking was missing in ∼15% of the study population. A sensitivity analysis with no imputation on smoking did not alter results (adjusted OR for major CHDs 1.35; 95% CI 1.30–1.40, P < .001 for all countries combined, singletons and multiples).

In a second sensitivity analysis, we excluded observations from Norway and added an adjustment for maternal highest educational level, data which were not available for Norway. Including singletons and multiples from Denmark, Finland, and Sweden, the adjusted OR for major CHDs was 1.36 (95% CI 1.30–1.41, P < .001).

The analysis including Finnish data with validated major CHDs showed similar results as the main analysis (adjusted OR for major CHDs 1.35; 95% CI 1.30–1.41, P < .001 for all countries combined, and adjusted OR 1.31; 95% CI 1.19–1.44, P < .001 for Finland).

Including 2 783 464 infants born between 2006 and 2015, small differences in rates of any major CHDs and severe CHDs were found, compared with the whole time period (see Supplementary data online, Table S4). For major CHDs, all countries combined, the absolute rates during this period were 1.88% for ART and 1.42% for SC and 0.34% and 0.25% for severe CHDs, respectively. For Denmark, the absolute rates for both major CHDs and severe CHDs decreased, while for the other Nordic countries, the rates during the more recent time period stayed almost unchanged or varied slightly up or down. The adjusted ORs remain, however, rather unchanged (major CHDs, adjusted OR 1.31, 95% CI 1.24–1.37, P < .001; severe CHDs, adjusted OR 1.35, 95% CI 1.20–1.52, P < .001). Lastly, including only Swedish data and adding paternal CHDs as a covariate did not change the results for major CHDs (adjusted OR 1.22; 95% CI 1.16–1.29, P < .001).

Congenital heart defects according to the classification of Botto

According to the hierarchical CHD classification, the risk of CHDs was higher in children born after ART than in spontaneously conceived children for five of the six lesion groups: conotruncal defects, non-conotruncal defects, VSD, ASD, and other CHDs (Table 2).

Singletons born after ART had increased risk for four lesion groups compared with spontaneously conceived singletons: conotruncal defects, VSD, ASD, and other CHDs (Table 3).

For multiples born after ART vs. singletons born after ART a higher risk was seen for five of the six lesion groups: non-conotruncal defects, coarctation aortae, VSD, ASD, and other CHDs, and for multiples born after SC vs. spontaneously conceived singletons, the risk was increased for all six lesion groups (Table 4).

In singletons born after ART, no difference was seen between ICSI and IVF (Table 5), or between frozen and fresh embryo transfer (Table 6) for any of the six lesion groups.

Selected specific congenital heart defects

We analysed 10 selected major CHDs (Table 7). In singletons, significantly increased risks in ART were seen for three CHDs: isomerism of atrial appendages, atrioventricular septal defects, and tetralogy of Fallot. Also including multiples, increased risk was seen also for pulmonary valve atresia (see Supplementary data online, Table S5).

Table 7

Risk of selected major congenital heart defects in singletons conceived by assisted reproductive technology vs. spontaneous conception (Denmark 1994–2014, Finland 1990–2014, Norway 1984–2015, and Sweden 1987–2015)

	No. of children		Risk of CHD, ART vs. spontaneous conception
	ART n = 127 275	SC n = 7 380 916	Crude OR (95% CI) P value	Adjusted ORa (95% CI) P value
Common arterial truncus (Q20.0b), n (%)	15 (0.01)	613 (0.01)	1.42 (0.85–2.37) .180	1.50 (0.89–2.52) .126
Double outlet right ventricle (Q20.1b), n (%)	23 (0.02)	831 (0.01)	1.61 (1.06–2.43) .025	1.26 (0.83–1.91) .281
Complete transposition of the great vessel (Q20.3b), n (%)	46 (0.04)	2676 (0.04)	1.00 (0.74–1.33) .983	0.99 (0.74–1.33) .949
Isomerism of atrial appendages with asplenia or polysplenia (Q20.6b), n (%)	11 (0.01)	196 (0.003)	3.25 (1.77–5.97) < .001	2.79 (1.50–5.18) < .001
Atrioventricular septal defect (Q21.2b), n (%)	90 (0.07)	3327 (0.05)	1.57 (1.27–1.93) < .001	1.28 (1.03–1.58) .023
Tetralogy of Fallot (Q21.3b), n (%)	60 (0.05)	2272 (0.03)	1.53 (1.19–1.98) < .001	1.34 (1.03–1.73) .028
Pulmonary valve atresia (Q22.0b), n (%)	29 (0.02)	1407 (0.02)	1.20 (0.83–1.73) .342	1.45 (1.00–2.10) .052
Tricuspid atresia and stenosis (Q22.4b), n (%)	9 (0.01)	447 (0.01)	1.17 (0.60–2.26) .645	1.37 (0.70–2.67) .357
Hypoplastic left heart syndrome (Q23.4b), n (%)	26 (0.02)	1607 (0.02)	0.94 (.64–1.38) .747	1.05 (0.71–1.54) .825
Coarctation aortae (Q25.1b), n (%)	80 (0.06)	4112 (0.06)	1.13 (0.90–1.41) .285	0.97 (0.78–1.22) .808

ART, assisted reproductive technology; CI, confidence interval; OR, odds ratio.

^aAdjustment for child’s year of birth, country of birth, maternal age.

^bICD-10 codes. Corresponding included ICD-8 and ICD-9 codes are shown in Supplementary data online, Table S2. A child can have more than one CHD diagnosis.

Discussion

In this large cohort study of 7.7 million liveborn children, including more than 171 000 children born after ART, we found that ART was associated with an increased risk of major CHDs as well as severe CHDs in both the overall ART population and in the ART singleton population, compared with spontaneously conceived children. Multiples, regardless of conception method, were associated with the highest risk of CHDs. Similar risks were observed in multiples conceived by ART and spontaneous conception, but this comparison is limited by the fact that we were missing information about chorionicity. The lower rate of monochorionic multiples in ART may give a false low risk in ART. Children conceived with ICSI did not seem to have an increased risk for CHDs compared with children conceived with IVF, and no significant difference was found between fresh and FET (Structured Graphical Abstract). The estimates were robust without any major changes after adjustments for available confounders or in sensitivity analyses.

Consistent with previous studies, our data showed higher occurrence of CHDs in pregnancies conceived by ART compared with spontaneously conceived pregnancies.^21–23 For specific CHDs, conflicting results have been reported. A large US study, including more than 11 million live births (singletons and multiples), of which 71 050 were conceived by ART, found a nearly three-fold increased risk of cyanotic CHDs in children born after ART, compared with children born after spontaneous conception, in adjusted analysis.⁴⁷ In a meta-analysis from 2018, Giorgione et al. analysed some specific CHDs in ART and spontaneously conceived singletons and multiples. They found lower occurrence of tetralogy of Fallot and transposition of the great arteries in the ART group. However, results were based on few events in the ART group.²¹ In contrast, a French case–control study of 1583 CHD cases and 4104 controls (singletons and multiples) assessing four different major structural CHDs found 2.4-fold odds of tetralogy of Fallot in children born after ART.²⁷

The overall risk of birth defects in our cohort has been explored in a previous study.¹¹ The study showed an increased risk of major birth defects in singletons conceived using ICSI with fresh embryo transfer compared with spontaneously conceived singletons. The risk was increased for most organ systems including the heart. Detailed data on type of CHD group or specific diagnoses were not reported. Further, multifetal pregnancies, an important mediator in risk of CHDs, were not included in our previous study.

Congenital heart defects are a heterogeneous group of diseases including both severe, life-threatening defects and minor abnormalities.^34,48,49 While most children with CHDs survive to adulthood, health issues persist for many children with CHDs when they grow up.^50,51 Children and adolescents with CHDs have an 11-fold increased risk of ischaemic stroke, compared with the general population, although absolute risk is low.^48,52 For adults with CHDs, the risks of pulmonary arterial hypertension and endocarditis are increased.^53,54 Further, for young adults with CHDs, 1 in 12 develop atrial fibrillation, and 1 in 10 of these develop congestive heart failure, before 42 years of age.^52,55

The aetiology of CHDs is mainly unknown, but chromosomal abnormalities and other genetic and environmental factors are considered to predispose to CHDs.^36,56 Congenital heart defects may be part of a malformation syndrome due to chromosomal aneuploidy, such as Down syndrome (trisomy 21), Edward syndrome (trisomy 18), Patau syndrome (trisomy 13), Turner syndrome (monosomy X), and Klinefelter syndrome (XXY), or Mendelian syndromes as Alagille–Holt–Oram syndrome and Noonan syndrome. Several environmental risk factors have been identified for CHDs, including both young and advanced maternal age, high parity, smoking, obesity, maternal diabetes, and use of drugs during pregnancy, e.g. antiepileptic and antidepressant drugs.^{38,41–44,57–61} Furthermore, women with a history of CHDs are considered to be at increased risk of having offspring with CHDs.^38,39,62,63 Also, low socioeconomic status has been found to be associated with CHDs.^64,65

Prenatal screening with foetal echocardiography for CHDs has been proposed to be beneficial for ART-conceived pregnancies.^21,66,67 and screening by foetal echocardiography is recommended by the American Heart Association for ART pregnancies.⁶⁶ Improved detection rate prenatally may offer the possibility for foetal therapy and/or specialized planning of delivery. However, this screening is still controversial and may cause increased costs and anxiety for the parents.^68–70 Further research is required to determine whether screening with foetal echocardiography, in addition to routine prenatal screening, will reduce morbidity and mortality for ART-conceived children when a major CHD is detected prenatally. In addition, preimplantation genetic testing may identify some CHDs of genetic origin and thereby contribute to decrease the CHDs among liveborn children.

Recent research has hypothesized that the placenta has a role in the development of CHDs, since placental vascular resistance has a direct impact on foetal circulation and thereby the developing foetal heart.⁷¹ Children born with CHDs have smaller placentas with increased vascular abnormalities.⁷² Further, studies also show a strong association between preeclampsia and CHDs, especially in early-onset and severe preeclampsia.^73,74 Pregnancies conceived with ART, in particular after FET, are associated with increased risk of preeclampsia, both for singleton and multifetal pregnancies.^75–77 An association between preeclampsia and CHD would, however, be expected to translate into a higher risk of CHD after FET which was not observed in the present study.

Twin pregnancies, especially monochorionic twins are associated with a higher risk of CHDs.⁷⁸ In recent years, multi-foetal pregnancies in ART have been declining, due to the introduction of the single embryo transfer policy.⁷⁹ However, the incidence of twin pregnancies continues to be elevated in ART-conceived pregnancies.^3,80

The main strength of this study is the large population with pooled nationwide data cross-linked from several high-quality national registries. Moreover, we explored specific CHD groups and specific assisted reproductive techniques. Detailed information enabled sub-analysis and adjustment for several confounders and comparisons according to multiplicity.

Some limitations should be considered when interpreting the results. Despite similar demography and healthcare systems, the rate of CHDs varied somewhat between countries. The follow-up for Norway was limited to birth, explaining the lower rate of CHDs in Norway. The reason for discrepancies between the other Nordic countries is not known but may be due to differences in registration policies and screening for foetal anomalies. The detection rate of major CHDs prenatally has increased substantially over time, as shown in Denmark leading to an increased termination of pregnancies, with a subsequent decrease in live-birth incidence of major CHDs.¹⁸ This change might have had an impact on the results in this study, particularly since a greater proportion of the ART cohort are born in later years in this study. There were some differences in prenatal screening routines in the four Nordic countries during the study period. All countries had introduced a second trimester ultrasound (gestational week 18–21) between 2004–07 including foetal organ screening and where the large majority of women participated. Norway had a second trimester prenatal screening ultrasound during the whole study period. A first-trimester ultrasound to assess the nuchal fold and determine the risk of aneuploidy was more variably introduced with a higher frequency in Denmark and Finland. Although sensitivity analyses showed some differences in rates of major and severe CHDs in live births in the later years, this seemed to occur in similar way for both ART and spontaneous conception, resulting in only minor changes in adjusted ORs. However, still these changes over time in combination with the much increasing ART population are considered a limitation.

Furthermore, a limitation of this study is the lack of information on CHDs in miscarriages, termination of pregnancies, and stillbirths. This may result in bias if ART-conceived pregnancies have a different probability of prenatal diagnosis with subsequent termination compared with spontaneously conceived pregnancies. A French study by Tararbit et al.⁸¹ found however no difference between ART and SC when evaluating the probability of prenatal diagnosis or termination of pregnancy for CHDs. A previous study on singletons from our cohort indicated similar risk of stillbirth after fresh and frozen embryo transfer compared with singletons conceived without medical assistance.⁸² One study limitation is that we relied only on registry data and ICD codes with the potential for miscoding. Some ICD codes, e.g. the codes for VSDs, do not differentiate between severe and less severe CHDs, and we should have needed more data on echocardiography and surgical and other procedures for correct classification. However, we have used different classifications of major CHDs to identify the most complex CHDs. Furthermore, a sensitivity analysis using data from the Finnish birth defects registry with validated major CHDs showed similar results as the main result.

Children conceived after ovulation induction and intrauterine insemination were included in the SC group. This misclassification will, if anything, dilute the association between ART and CHD.¹² Other limitations are that we did not have information about causes of infertility and data on specific techniques used in assisted reproduction was not available from Finland. Finally, as in all observational studies residual confounding by unknown or unmeasured factors may remain.

Conclusions

Congenital heart defects are serious, although rare conditions. This large study reports a higher occurrence of CHDs after ART conception, both severe and less severe. The highest rates of CHDs were observed in children born in multiple pregnancies. No difference in CHDs was found between ICSI and IVF and neither between children born after fresh or frozen transfer. The findings of the current study should be conveyed to patients undergoing counselling before ART. Although the risk for major CHDs is higher in children born after ART, the absolute increase in risks seems to be modest. This study also emphasizes the importance of single embryo transfer to avoid the increased risks in multifetal pregnancies.

Acknowledgements

We thank all national registries providing individual data for this study.

Supplementary data

Supplementary data are available atInterest European Heart Journal online.

Declarations

Disclosure of Interest

All authors declare no disclosure of interest for this contribution.

Data Availability

The data that support the findings of this study are not available due to regulations restricted by law. Storage of data is arranged and ensured by Statistics Denmark. Research was feasible after receiving approvals from the Ethics Committees and registry-keeping authorities in each country. Administration of microdata from the national registries is regulated by the General Data Protection Regulation (GDPR) ensured by Statistics Denmark. Contact information for Statistics Denmark: Division of Research Services Statistics Denmark Sejrøgade 11 DK-2100 Copenhagen Denmark E-mail: [email protected] Phone: + 45 39 17 31 30.

Funding

The CoNARTaS has been supported by the Nordic Trial Alliance: a pilot project jointly funded by the Nordic Council of Ministers and NordForsk (https://www.nordforsk.org/sv/research-areas/nordic-trial-alliance) (grant number 71450; A.P.), the Central Norway Regional Health Authorities (grant number 46045000), the Norwegian Cancer Society (grant number 182356-2016; S.O.], the Nordic Federation of Obstetrics and Gynaecology [grant numbers NF13041, NF15058, NF16026, NF17043, and NF399 (2022–23)], the Interreg Öresund-Kattegat-Skagerrak European Regional Development Fund (ReproUnion project; AP), and the Swedish state under the agreement between the Swedish government and the county councils, the ALF-agreement (ALFGBG-70940; C.B.), and the Hjalmar Svensson Foundation (U.-B.W.). The funding sources had no role in study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the article for publication.

Ethical Approval

Ethical approvals were obtained from Ethical Committee in Gothenburg, Sweden (Dnr 214-12, T422-12, T516-15, T233-16, T300-17, T1144-17, T121-18, T1071-18, 2019-02347, 2022-00903-02), and in Norway from the Regional Committee for Medical and Health Research Ethics (REK-Nord, 2010/1909). There are no requirements for ethical approval for registry-based studies in Denmark and Finland.

Pre-registered Clinical Trial Number

ISRCTN 11780826.

Author notes

Christina Bergh and Ulla-Britt Wennerholm contributed equally to the study.