[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure.  Framework of the Association Between Assisted Reproductive Technology and Congenital Heart Defects With Twin Pregnancy as the Mediator
Framework of the Association Between Assisted Reproductive Technology and Congenital Heart Defects With Twin Pregnancy as the Mediator
Table 1.  Demographic Characteristics and Medical History of Women With ART and Non-ART Pregnanciesa
Demographic Characteristics and Medical History of Women With ART and Non-ART Pregnanciesa
Table 2.  Association of Congenital Heart Defects With Potential Confounding Risk Factors Among Non-ART Pregnanciesa
Association of Congenital Heart Defects With Potential Confounding Risk Factors Among Non-ART Pregnanciesa
Table 3.  Association of Congenital Heart Defects With ART and Twin Status
Association of Congenital Heart Defects With ART and Twin Status
Table 4.  Analysis of ART and Congenital Heart Defects Association With Twin Pregnancy as a Mediatora
Analysis of ART and Congenital Heart Defects Association With Twin Pregnancy as a Mediatora
1.
Allen  VM, Wilson  RD, Cheung  A; Genetics Committee; Reproductive Endocrinology and Infertility Committee.  Pregnancy outcomes after assisted reproductive technology.   J Obstet Gynaecol Can. 2006;28(3):220-233. doi:10.1016/S1701-2163(16)32112-0 PubMedGoogle ScholarCrossref
2.
Toner  JP, Coddington  CC, Doody  K,  et al.  Society for Assisted Reproductive Technology and assisted reproductive technology in the United States: a 2016 update.   Fertil Steril. 2016;106(3):541-546. doi:10.1016/j.fertnstert.2016.05.026 PubMedGoogle ScholarCrossref
3.
Bahtiyar  MO, Campbell  K, Dulay  AT,  et al.  Is the rate of congenital heart defects detected by fetal echocardiography among pregnancies conceived by in vitro fertilization really increased? a case-historical control study.   J Ultrasound Med. 2010;29(6):917-922. doi:10.7863/jum.2010.29.6.917 PubMedGoogle ScholarCrossref
4.
Giorgione  V, Parazzini  F, Fesslova  V,  et al.  Congenital heart defects in IVF/ICSI pregnancy: systematic review and meta-analysis.   Ultrasound Obstet Gynecol. 2018;51(1):33-42. doi:10.1002/uog.18932 PubMedGoogle ScholarCrossref
5.
Hansen  M, Kurinczuk  JJ, Bower  C, Webb  S.  The risk of major birth defects after intracytoplasmic sperm injection and in vitro fertilization.   N Engl J Med. 2002;346(10):725-730. doi:10.1056/NEJMoa010035 PubMedGoogle ScholarCrossref
6.
Hansen  M, Bower  C, Milne  E, de Klerk  N, Kurinczuk  JJ.  Assisted reproductive technologies and the risk of birth defects—a systematic review.   Hum Reprod. 2005;20(2):328-338. doi:10.1093/humrep/deh593 PubMedGoogle ScholarCrossref
7.
Hansen  M, Kurinczuk  JJ, Milne  E, de Klerk  N, Bower  C.  Assisted reproductive technology and birth defects: a systematic review and meta-analysis.   Hum Reprod Update. 2013;19(4):330-353. doi:10.1093/humupd/dmt006 PubMedGoogle ScholarCrossref
8.
Lancaster  PA.  Congenital malformations after in-vitro fertilisation.   Lancet. 1987;2(8572):1392-1393. doi:10.1016/S0140-6736(87)91277-3 PubMedGoogle ScholarCrossref
9.
Wen  SW, Leader  A, White  RR,  et al.  A comprehensive assessment of outcomes in pregnancies conceived by in vitro fertilization/intracytoplasmic sperm injection.   Eur J Obstet Gynecol Reprod Biol. 2010;150(2):160-165. doi:10.1016/j.ejogrb.2010.02.028 PubMedGoogle ScholarCrossref
10.
Panagiotopoulou  O, Fouzas  S, Sinopidis  X, Mantagos  SP, Dimitriou  G, Karatza  AA.  Congenital heart disease in twins: the contribution of type of conception and chorionicity.   Int J Cardiol. 2016;218:144-149. doi:10.1016/j.ijcard.2016.05.029 PubMedGoogle ScholarCrossref
11.
Votava-Smith  JK, Glickstein  JS, Simpson  LL, Williams  IA.  Comparison of method of conception in fetuses undergoing echocardiography at a tertiary referral center.   Prenat Diagn. 2014;34(5):445-449. doi:10.1002/pd.4327 PubMedGoogle ScholarCrossref
12.
Tararbit  K, Houyel  L, Bonnet  D,  et al.  Risk of congenital heart defects associated with assisted reproductive technologies: a population-based evaluation.   Eur Heart J. 2011;32(4):500-508. doi:10.1093/eurheartj/ehq440 PubMedGoogle ScholarCrossref
13.
Tararbit  K, Lelong  N, Thieulin  AC,  et al; EPICARD Study Group.  The risk for four specific congenital heart defects associated with assisted reproductive techniques: a population-based evaluation.   Hum Reprod. 2013;28(2):367-374. doi:10.1093/humrep/des400 PubMedGoogle ScholarCrossref
14.
Tararbit  K, Lelong  N, Houyel  L, Bonnet  D, Goffinet  F, Khoshnood  B; EPICARD Study Group.  Assessing the role of multiple pregnancies in the association between tetralogy of Fallot and assisted reproductive techniques: a path-analysis approach.   Orphanet J Rare Dis. 2014;9:27. doi:10.1186/1750-1172-9-27 PubMedGoogle ScholarCrossref
15.
Davies  MJ, Moore  VM, Willson  KJ,  et al.  Reproductive technologies and the risk of birth defects.   N Engl J Med. 2012;366(19):1803-1813. doi:10.1056/NEJMoa1008095 PubMedGoogle ScholarCrossref
16.
Simeone  RM, Tinker  SC, Gilboa  SM,  et al; National Birth Defects Prevention Study.  Proportion of selected congenital heart defects attributable to recognized risk factors.   Ann Epidemiol. 2016;26(12):838-845. doi:10.1016/j.annepidem.2016.10.003 PubMedGoogle ScholarCrossref
17.
Cedergren  MI, Källén  BA.  Maternal obesity and infant heart defects.   Obes Res. 2003;11(9):1065-1071. doi:10.1038/oby.2003.146 PubMedGoogle ScholarCrossref
18.
Pasquali  R, Patton  L, Gambineri  A.  Obesity and infertility.   Curr Opin Endocrinol Diabetes Obes. 2007;14(6):482-487. doi:10.1097/MED.0b013e3282f1d6cb PubMedGoogle ScholarCrossref
19.
Hur  YM, Song  TB.  A recent rise in twin birth rates and demographic changes in mothers of twins in South Korea: 2003-2007.   Twin Res Hum Genet. 2009;12(1):118-122. doi:10.1375/twin.12.1.118 PubMedGoogle ScholarCrossref
20.
Kurosawa  K, Masuno  M, Kuroki  Y.  Trends in occurrence of twin births in Japan.   Am J Med Genet A. 2012;158A(1):75-77. doi:10.1002/ajmg.a.34362 PubMedGoogle ScholarCrossref
21.
Dawson  AL, Tinker  SC, Jamieson  DJ,  et al; National Birth Defects Prevention Study.  Twinning and major birth defects, National Birth Defects Prevention Study, 1997-2007.   J Epidemiol Community Health. 2016;70(11):1114-1121. doi:10.1136/jech-2015-206302 PubMedGoogle ScholarCrossref
22.
Huang  CT, Au  HK, Chien  LW, Chang  CW, Chien  YY, Tzeng  CR.  Twin pregnancy outcome among cases of spontaneous conception, intrauterine insemination, and in vitro fertilization/intracytoplasmic sperm injection.   Fertil Steril. 2006;86(4):1017-1019. doi:10.1016/j.fertnstert.2006.02.109 PubMedGoogle ScholarCrossref
23.
Thurin  A, Hausken  J, Hillensjö  T,  et al.  Elective single-embryo transfer versus double-embryo transfer in in vitro fertilization.   N Engl J Med. 2004;351(23):2392-2402. doi:10.1056/NEJMoa041032 PubMedGoogle ScholarCrossref
24.
Koudstaal  J, Bruinse  HW, Helmerhorst  FM, Vermeiden  JP, Willemsen  WN, Visser  GH.  Obstetric outcome of twin pregnancies after in-vitro fertilization: a matched control study in four Dutch university hospitals.   Hum Reprod. 2000;15(4):935-940. doi:10.1093/humrep/15.4.935 PubMedGoogle ScholarCrossref
25.
Pharoah  PO.  Twins and cerebral palsy.   Acta Paediatr Suppl. 2001;90(436):6-10. doi:10.1111/j.1651-2227.2001.tb01620.x PubMedGoogle ScholarCrossref
26.
Hardin  J, Carmichael  SL, Selvin  S, Lammer  EJ, Shaw  GM.  Increased prevalence of cardiovascular defects among 56,709 California twin pairs.   Am J Med Genet A. 2009;149A(5):877-886. doi:10.1002/ajmg.a.32745 PubMedGoogle ScholarCrossref
27.
Bernier  PL, Stefanescu  A, Samoukovic  G, Tchervenkov  CI.  The challenge of congenital heart disease worldwide: epidemiologic and demographic facts.   Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu. 2010;13(1):26-34. doi:10.1053/j.pcsu.2010.02.005 PubMedGoogle ScholarCrossref
28.
Canobbio  MM.  Health care issues facing adolescents with congenital heart disease.   J Pediatr Nurs. 2001;16(5):363-370. doi:10.1053/jpdn.2001.26570 PubMedGoogle ScholarCrossref
29.
Lui  GK, Rogers  IS, Ding  VY,  et al.  Risk estimates for atherosclerotic cardiovascular disease in adults with congenital heart disease.   Am J Cardiol. 2017;119(1):112-118. doi:10.1016/j.amjcard.2016.09.023 PubMedGoogle ScholarCrossref
30.
Jay  PY, Akhirome  E, Magnan  RA,  et al.  Transgenerational cardiology: one way to a baby’s heart is through the mother.   Mol Cell Endocrinol. 2016;435:94-102. doi:10.1016/j.mce.2016.08.029 PubMedGoogle ScholarCrossref
31.
Kloesel  B, DiNardo  JA, Body  SC.  Cardiac embryology and molecular mechanisms of congenital heart disease: a primer for anesthesiologists.   Anesth Analg. 2016;123(3):551-569. doi:10.1213/ANE.0000000000001451 PubMedGoogle ScholarCrossref
32.
American Institute of Ultrasound in Medicine.  AIUM practice parameter for the performance of fetal echocardiography.   J Ultrasound Med. 2020;39(1):E5-E16.PubMedGoogle Scholar
33.
Better Outcomes Registry & Network (BORN). Ottawa, Ontario, Canada. http://www.bornontario.ca. Accessed March 25, 2019.
34.
Centers for Disease Control and Prevention, US Department of Health & Human Services. Assisted reproductive technology (ART). https://www.cdc.gov/art/whatis.html. Accessed March 25, 2019.
35.
Zhai  D, Guo  Y, Smith  G, Krewski  D, Walker  M, Wen  SW.  Maternal exposure to moderate ambient carbon monoxide is associated with decreased risk of preeclampsia.   Am J Obstet Gynecol. 2012;207(1):57.e1-57.e9. doi:10.1016/j.ajog.2012.03.022 PubMedGoogle ScholarCrossref
36.
Wilkins  R, Khan  S. Automated geographic coding based on the statistics Canada postal code conversion files: including postal codes through October 2010. https://mdl.library.utoronto.ca/sites/default/files/mdldata/open/canada/national/statcan/postalcodes/pccfplus/2006/2010oct/MSWORD.PCCF5H.pdf. Accessed July 15, 2019.
37.
Hernán  MA, Hernández-Díaz  S, Werler  MM, Mitchell  AA.  Causal knowledge as a prerequisite for confounding evaluation: an application to birth defects epidemiology.   Am J Epidemiol. 2002;155(2):176-184. doi:10.1093/aje/155.2.176 PubMedGoogle ScholarCrossref
38.
Schisterman  EF, Cole  SR, Platt  RW.  Overadjustment bias and unnecessary adjustment in epidemiologic studies.   Epidemiology. 2009;20(4):488-495. doi:10.1097/EDE.0b013e3181a819a1 PubMedGoogle ScholarCrossref
39.
Rothman  KJ, Greenland  S, Lash  TL.  Modern Epidemiology. 3rd ed. New York, NY: Wolters Kluwer/Lippincott Williams & Wilkins; 2008.
40.
VanderWeele  TJ.  Explanation in Causal Inference: Methods for Mediation and Interaction. New York, NY: Oxford University Press, 2015.
41.
VanderWeele  TJ.  Mediation analysis: a practitioner’s guide.   Annu Rev Public Health. 2016;37:17-32. doi:10.1146/annurev-publhealth-032315-021402 PubMedGoogle ScholarCrossref
42.
VanderWeele  TJ.  A unification of mediation and interaction: a 4-way decomposition.   Epidemiology. 2014;25(5):749-761. doi:10.1097/EDE.0000000000000121 PubMedGoogle ScholarCrossref
43.
Richiardi  L, Bellocco  R, Zugna  D.  Mediation analysis in epidemiology: methods, interpretation and bias.   Int J Epidemiol. 2013;42(5):1511-1519. doi:10.1093/ije/dyt127 PubMedGoogle ScholarCrossref
44.
van Buuren  S.  Multiple imputation of discrete and continuous data by fully conditional specification.   Stat Methods Med Res. 2007;16(3):219-242. doi:10.1177/0962280206074463 PubMedGoogle ScholarCrossref
45.
Little  RJA, Rubin  DB.  Statistical Analysis With Missing Data. 2nd ed. New York, NY: John Wiley & Sons; 2002. doi:10.1002/9781119013563
46.
Rubin  DB.  Multiple Imputation for Nonresponse in Surveys. New York, NY: John Wiley & Sons Inc; 2004.
47.
Muzyka  D, Lypka  M. The benefits of keeping it casual: investigating stress mediated health outcomes with PROC CAUSALMED. https://www.sas.com/content/dam/SAS/support/en/sas-global-forum-proceedings/2019/3897-2019.pdf. Accessed July 15, 2019.
48.
Procrea Fertility. Ontario treatment coverage and payment options. https://procrea.ca/financial-guidance/ontario-treatment-coverage-payment-options/ Accessed September 3, 2019.
49.
Oberg  AS, VanderWeele  TJ, Almqvist  C, Hernandez-Diaz  S.  Pregnancy complications following fertility treatment-disentangling the role of multiple gestation.   Int J Epidemiol. 2018;47(4):1333-1342. doi:10.1093/ije/dyy103 PubMedGoogle ScholarCrossref
50.
VanderWeele  TJ, Ding  P.  Sensitivity analysis in observational research: introducing the E-value.   Ann Intern Med. 2017;167(4):268-274. doi:10.7326/M16-2607 PubMedGoogle ScholarCrossref
51.
Better Outcomes Registry & Network. BORN data quality report, 2012-2014—executive summary. https://www.bornontario.ca/en/data/resources/Documents/BORN-Data-Quality-Report-2012-2014---Executive-Summary.pdf. Published April 2016. Accessed March 25, 2019.
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Views 2,714
    Citations 0
    Original Investigation
    February 24, 2020

    Associations of Assisted Reproductive Technology and Twin Pregnancy With Risk of Congenital Heart Defects

    Author Affiliations
    • 1OMNI Research Group, Ottawa Hospital Research Institute, Ottawa, Ontario, Canada
    • 2Department of Obstetrics and Gynecology, University of Ottawa Faculty of Medicine, Ottawa, Ontario, Canada
    • 3School of Epidemiology and Public Health, University of Ottawa Faculty of Medicine, Ottawa, Ontario, Canada
    • 4Nanhai Hospital, Southern Medical University, Foshan, Guangdong, China
    • 5Better Outcomes Registry & Network (BORN) Ontario, Children’s Hospital of Eastern Ontario, Ottawa, Ontario, Canada
    • 6Children’s Hospital of Eastern Ontario Research Institute, Ottawa, Ontario, Canada
    • 7Division of Cardiology, Children’s Hospital of Eastern Ontario, Ottawa, Ontario, Canada
    • 8Department of Pediatrics, University of Ottawa Faculty of Medicine, Ottawa, Ontario, Canada
    • 9Robinson Research Institute, University of Adelaide, Adelaide, South Australia, Australia
    JAMA Pediatr. 2020;174(5):446-454. doi:10.1001/jamapediatrics.2019.6096
    Key Points

    Question  Is assisted reproductive technology associated with increased risk of congenital heart defects independent of its association with twin pregnancies?

    Findings  In this cohort study of 507 390 singleton or twin pregnancies, the prevalence of congenital heart defects was higher in assisted pregnancies (223 [2.2%]) than in nonassisted pregnancies (6057 [1.2%]). Twinning mediated 87.3% of the association between assisted pregnancy and congenital heart defects.

    Meaning  The findings suggest the association between assisted reproductive technology and congenital heart defects may be mediated by twinning.

    Abstract

    Importance  The extent to which assisted reproductive technology is associated with increased risk of congenital heart defects independent of its known association with twinning remains uncertain.

    Objective  To assess the extent to which assisted pregnancy is associated with increased risk of congenital heart defects independent of its known association with twinning.

    Design, Setting, and Participants  This retrospective cohort study linked records of congenital heart defect diagnoses with assisted reproductive technology cycles in 507 390 singleton or twin pregnancies (10 149 assisted pregnancies and 497 241 nonassisted pregnancies), including singleton and twin early pregnancy losses, stillbirths, and live births (follow-up to 1 year of age) in Ontario, Canada, between April 1, 2012, and October 31, 2015. Statistical analysis was performed from January 1, 2017, to September 9, 2019.

    Exposures  Assisted reproductive technology and its 2 subtypes: intracytoplasmic sperm injection and in vitro fertilization without intracytoplasmic sperm injection.

    Main Outcomes and Measures  The main outcome was congenital heart defects (prevalence and relative risk measured as odds ratios [ORs]). Mediation analysis was performed to assess the extent to which the association between assisted reproductive technology and congenital heart defects was mediated by twinning.

    Results  Of 507 390 mother-infant pairs with singleton or twin pregnancies evaluated, the prevalence of congenital heart defects in assisted pregnancies (223 [2.2%]) was higher than that in nonassisted pregnancies (6057 [1.2%]; crude OR, 1.82; 95% CI, 1.59-2.09). The strength of the association between assisted pregnancy and congenital heart defects decreased after adjusting for several risk factors simultaneously (adjusted OR, 1.70; 95% CI, 1.48-1.95). Further mediation analysis indicated that most of the association between assisted pregnancy and congenital heart defects was mediated by twinning (adjusted OR, 1.68; 95% CI, 1.44-1.92), and the natural direct association of assisted pregnancy with congenital heart defects among singleton pregnancies was 1.09 (95% CI, 0.93-1.25). Mediation of twinning accounted for 87.3% of the association.

    Conclusions and Relevance  Our study results suggest that the association between assisted reproductive technology and congenital heart defects may be mediated by twinning.

    Introduction

    At present, approximately 2% of pregnancies are conceived by using assisted reproductive technology (ART), with upward trends in recent years.1,2 An increased risk of adverse pregnancy outcomes, including congenital heart defects (CHD), in ART pregnancies compared with non-ART pregnancies has been observed.1-15 Several maternal factors, such as maternal age, obesity, socioeconomic status, and preexisting health problems, have been associated with risks of both adverse pregnancy outcomes, including CHD and infertility requiring ART to conceive,1,2,16-18 and the association between ART and CHD tends to be reduced after adjustment for these maternal factors.3-15

    One important consequence of increased use of ART is a substantially increased rate of twin pregnancies.1,2,19-26 Compared with singletons, the risks of adverse pregnancy outcomes including CHD have been shown to be increased in twins.21,22,24-26 Some studies have suggested that part of the increased risk of CHD observed in ART pregnancies may be attributable to the increased number of twin pregnancies.10,14,26

    Congenital heart defect is a serious disease burden not only in infants but also in adolescents and adults.27-29 The pathophysiologic findings and pathogenesis of CHD are complicated and often are very different from other congenital anomalies.12,13,30,31 The American Institute of Ultrasound in Medicine, in collaboration with other professional organizations, has developed the Fetal Echocardiography Practice Parameter,32 and the American College of Radiology has endorsed this document. The association of ART and CHD should be further examined to decide whether fetal echocardiography, a costly procedure, should be used in ART pregnancies.32

    We performed a large population-based study to examine the association between ART and CHD and to explore the associations of ART, twin pregnancies, and CHD, with twin status considered as a mediator in the analysis.

    Methods
    Study Design and Study Population

    For this cohort study, ethical approval was obtained from the Ottawa Hospital Research Ethical Board, Ottawa, Ontario, Canada. Because this study used data collected by the province-approved registry with no additional data collection from patients, no informed consent was required. All singleton and twin stillbirths and live births (with follow-up to 1 year of age) and some early pregnancy losses that occurred between April 1, 2012, and October 31, 2015, in Ontario, Canada, and were captured by Better Outcomes Registry & Network (BORN) were included in the study. BORN is a provincial maternal child registry aiming to facilitate and improve care for mothers and children in Ontario.33 BORN has an internet-based data entry system administered by the Children’s Hospital of Eastern Ontario called the BORN Information System. There are 3 locations and periods in the BORN Information System in which CHD data can be captured: (1) prenatal screening records, (2) antenatal specialty clinic records, and (3) hospital records for birth or postpartum care or neonatal intensive care unit admissions. BORN has 100% capture for hospital and home births in Ontario. This database contains maternal demographic characteristics and health behaviors, preexisting maternal health problems, and birth outcomes. The national Canadian Assisted Reproductive Technologies Registry (CARTR Plus) is housed within BORN and is owned and administered by the medical directors of in vitro fertilization clinics and the Canadian Fertility and Andrology Society. The national CARTR Plus includes in vitro fertilization cycles for individuals undergoing ART-associated treatment in Canada.

    Outcome

    Congenital heart defect was the outcome of interest in this study. To reduce misclassifications, only major structural CHD was considered (for a list of major structural CHD, see eTable 1 in the Supplement). Newborn diagnoses including CHD were collected at birth and during the postpartum period. Information on congenital anomalies in lost or terminated pregnancies was captured from prenatal screening follow-up visits if the anomaly was diagnosed through prenatal screening and antenatal specialty clinics if the anomaly was diagnosed through an antenatal clinic in Ontario. For CHD cases that were not captured by BORN, an additional ascertainment occurred through linkage with the Canadian Institute for Health Information (CIHI) Discharge Abstract Database and National Ambulatory Care Reporting System metadata. In CIHI records, CHD cases were identified based on a standardized diagnostic coding system, International Classification of Diseases and Related Health Problems, Tenth Revision, Canada (ICD-10-CA). We were able to capture CHD at birth and up to 1 year through linkage with CIHI records. Because of difficulties in separately identifying in utero fetal CHD cases in twins and limited further linking to infant records, we used individual pregnancy as the unit of analysis (ie, if 1 member of a twin pair was affected, that pregnancy was considered to be affected).

    Exposure

    Assisted reproductive technology was the exposure of interest in the study. We followed the Centers for Disease Control and Prevention definition for ART,34 which included both intracytoplasmic sperm injection (ICSI) and in vitro fertilization (IVF) without ICSI. The report of treatment cycles by fertility centers is not uniformly required, and some cycles may be missed. Moreover, BORN started incorporating CARTR Plus data starting in January 2013. To maximize ascertainment of ART, we used BORN records on conception type to supplement CARTR Plus records.

    Confounders

    Potential confounding variables accounted for were maternal age, parity, prepregnancy obesity (defined by prepregnancy body mass index), maternal smoking, social drug use or alcohol consumption, folic acid intake during pregnancy, mental health illness during prepregnancy or pregnancy (a composite measure of psychiatric disorders, depression, and anxiety), and prepregnancy physical health problems (a composite measure of chronic hypertension, diabetes, heart disease, thyroid disease, lupus, alcoholism, asthma, HIV infection, and hepatitis B). We also attempted to account for maternal educational level and family income. Because no individual records on maternal educational level and family income are available in the BORN Information System, we created 2 neighborhood-level variables for Ontario by converting postal codes affiliated with the mother into dissemination areas from the Canada census data nearest to the study period: (1) percentage of having a university degree among the adult population aged 25 to 64 years and (2) median family income, both measured in quintiles according to methods established in an earlier study.35 Dissemination areas were small, relatively stable geographic units with a population of 400 to 700 persons of relative homogenous socioeconomic status.36 We were careful in selecting confounders to be adjusted in the analysis so that no overadjustment would occur.37,38 All confounders included in the analysis were considered to be independently associated with CHD (outcome) and ART (exposure), and we verified the potential confounders by examining the association between confounders and outcome among unexposed study participants only.39

    Mediator

    Twins have remarkably higher risks of adverse perinatal outcomes including CHD,21,22,24-26 and the proportion of twins in ART pregnancies is substantially higher than natural pregnancies.1,2,19-23 Because twin status is the intermediate step between ART and CHD (twin status is determined before the development of CHD), it qualified as a potential mediator.40-43

    The pregnancy identification (an internal identifier assigned by BORN) was used to link records of the infant with the same pregnancy to obtain maternal information, and the birth identification (an internal identifier assigned by BORN) was used to link the records of the same infant across different periods within the BORN Information System. Unencrypted health card numbers were used to link BORN Information System records with CIHI records. The record linkage process is shown in the eFigure in the Supplement.

    Statistical Analysis

    Statistical analysis was performed from January 1, 2017, to September 9, 2019. All analyses were performed using SAS, version 9.4 (SAS Institute Inc). We first compared baseline characteristics between ART and non-ART pregnancies and between the 2 subgroups of ART. We then described the prevalence of CHD as associated with potential confounding risk factors. Before multivariable regression analysis, multiple imputation using a fully conditional specification method44-46 was applied to create complete data sets. Details of missing imputation are provided in the eMethods in the Supplement. Multivariable logistic regression analyses were then performed in both exposure-outcome and mediator-outcome models while accounting for covariates. The variable ART was assessed as a binary variable. Because CHD is a rare outcome (<5%), we used odds ratios (ORs) to estimate relative risk in logistic regression models. All prespecified risk factors were included in the analysis.

    To assess potential mediation of twinning in the association between ART and CHD, the Causalmed procedure47 in SAS was performed. The specific issue to be addressed by the mediation analysis was (1) how much of the overall association of ART with CHD risk was attributable to the association of ART with increased risk of twinning, which in turn affects CHD risk, and (2) how much of the association of ART and CHD did not involve twinning.40-43 The mediation analysis was done to parse the total association of ART with CHD prevalence into 2 components. The natural direct association of ART with CHD risk in the study population would occur if the pathway from ART to twin status were disabled. The Figure gives the contribution of the ART-CHD pathway to the overall association of ART with CHD risk. The natural indirect association, which is the association of ART with CHD risk in the study population, would occur if the direct ART-CHD pathway were disabled so that ART was associated with CHD risk only via the ART-twin-CHD pathway. The proportion mediated was defined as the natural indirect association divided by the total association, although it is only meaningful when both components of the total association have the same mathematical sign. Mediation analysis allows for control of confounding factors in estimating these results.

    Four-way decompositions of the total excess association were further estimated considering the potential interaction of ART with twinning in the association with risk of CHD using risk difference.42 The question of 4-way decompositions analysis is the following: To what extent is ART associated with CHD risk differently in twin pregnancies than in singleton pregnancies? The ART twin-status interaction can be operative regardless of the relative contributions of the direct and indirect paths (Figure). The ART twin-status interaction could be important for 2 reasons. First, it could help toward understanding the biologic mechanisms that underlie the association between ART and CHD. Why might ART affect twin pregnancies differently from how it affects singleton pregnancies in terms of CHD risk? Second, the amount of ART-twin interaction affects the degree to which results of mediation analysis obtained from this study are generalizable to other populations. The first of 4 components assessed in the 4-way decompositions was the controlled direct association, which was the CHD risk difference between ART and non-ART pregnancies among singleton births (the baseline risk difference in the absence of mediation or interaction). The second was reference interaction, which was the contribution of the ART-twin interaction if the frequency of twinning in ART pregnancies was set to what was observed among non-ART pregnancies (the association between CHD risk and the ART-twin interaction in the absence of mediation). The third component was the mediated interaction, which was the contribution of the ART-twin interaction when applied only to the added increase in the frequency of twinning seen in ART pregnancies (the combined association between CHD risk and mediation and the ART-twin interaction). The fourth component was the pure indirect association, which was the increase in CHD risk associated with the increased proportion of twins seen in ART pregnancies apart from the ART-twin interaction (mediation by twin status in the absence of an ART-twin interaction).

    Preliminary analysis showed that there was no major difference between IVF and ICSI in terms of their association with CHD. To preserve the statistical precision and to simplify the interpretation, mediation analysis was performed for overall ART only. To assess robustness of the results, analyses deleting missing data without missing imputation and CHDs ascertained at birth or up to 1 year after birth (ie, without linking CHDs with fetal losses or terminations) were performed.

    Results
    Baseline Characteristics

    A total of 507 390 mother-infant pairs with a singleton or twin birth were matched and included in final analysis. Among these, 10 149 (2.0%) were assisted pregnancies and 497 241 (98.0%) were nonassisted pregnancies. Women who conceived by ART were more likely to be older, to be experiencing their first pregnancy, and to have a higher body mass index. They were more likely to live in higher-income and higher-educational-level neighborhoods and have physical health problems but were less likely to have major mental health problems (such as schizophrenia or bipolar disease) and less likely to engage in smoking, social drug use, or alcohol consumption during pregnancy compared with women who conceived without ART (Table 1). The twin rate differed between groups: 18.5% (1881 of 10 149) in assisted pregnancies and 1.4% (7082 of 497 241) in nonassisted pregnancies. Women who conceived by ICSI were more likely to experience their first pregnancy (73.7% [2825 of 3835] vs 60.0% [3655 of 6094]) but less likely to be obese (16.8% [534 of 3187] vs 19.2% [945 of 4929]) than women who conceived by IVF. Twin rates were 17.5% in pregnancies conceived by ICSI and 19.2% in pregnancies conceived by IVF.

    Association of CHD With Potential Confounding Risk Factors

    Table 2 shows the association of CHD found in infants with potential maternal confounding risk factors. The prevalence of CHD was higher among infants born to older mothers (adjusted OR, 1.40 [95% CI, 1.31-1.51] for women >35 years vs <30 years); lower-income mothers (adjusted OR, 1.10 [95% CI, 1.00-1.20] for women of family income at 40%-59% levels; 1.06 [95% CI, 0.96-1.16] at 20%-39% levels; and 1.12 [95% CI, 1.02-1.22] at <20% levels compared with women of family income >80% level); mothers residing in lower-educational-level neighborhoods (adjusted OR, 1.13 [95% CI, 1.04-1.24] for women residing in neighborhoods at 60%-80% levels of university degree; 1.19 [95% CI, 1.09-1.31] at 40%-59% levels; 1.27 [95% CI, 1.16-1.40] at 20%-39% levels; and 1.28 [95% CI, 1.16-1.41] at <20% levels compared with women residing in >80% level neighborhoods); those who smoked or used drugs or consumed alcohol (adjusted OR, 1.39 [95% CI, 1.30-1.50] compared with those who did not); those who did not receive folic acid during pregnancy (adjusted OR, 1.08 [95% CI, 1.02-1.14] compared with those who did); those who had major physical problems (adjusted OR, 1.77 [95% CI, 1.67-1.88] compared with those who did not); those who had major mental health problems (adjusted OR, 1.26 [95% CI, 1.18-1.35] compared with those who did not); and those with obesity (adjusted OR, 1.21 [95% CI, 1.13-1.30] compared with those who were nonobese).

    Association of CHD With ART and Twin Pregnancies

    Table 3 displays the results of the analysis of the association of CHD with ART and twinning and ART-twin interactions. Infants born to mothers who conceived by ART had a higher prevalence of CHD than did infants born to mothers who conceived without using ART (2.2% [223 of 10 149] vs 1.2% [6057 of 497 241]; crude OR, 1.82; 95% CI, 1.59-2.09). This association decreased after adjusting for several risk factors simultaneously. For specific ART techniques, infants born to mothers who conceived by ICSI had higher risk of CHD than did infants born to mothers who conceived by IVF (2.3% vs 2.1%). Twins had a higher prevalence of CHD than singletons (5.5% vs 1.2%; OR, 4.91; 95% CI, 4.47-5.40). This association decreased after adjusting for several risk factors simultaneously. Prevalence of CHD was the highest in non-ART twin pregnancies (5.6%), second highest in ART twin pregnancies (5.1%), second lowest in ART singleton pregnancies (1.6%), and lowest in non-ART singleton pregnancies (1.2%).

    Mediation Analysis

    Table 4 presents results of mediation analysis with twinning acting as the mediator between ART and CHD. The results showed that the association between ART and CHD was largely mediated by twinning (adjusted OR, 1.68; 95% CI, 1.44-1.92). The OR of the natural direct association of ART with CHD among singleton pregnancies was 1.09 (95% CI, 0.93-1.25).

    The results of 4-way decompositions analysis found that reference interaction and mediated interaction together accounted for 7.9% of the total association, and neither component was statistically significantly different from 0. Mediation through twinning accounted for more than 80% of the excess association (eTable 2 in the Supplement). Results of multiple logistical regression analysis deleting missing data or restricting to pregnancy outcome yielded similar results in terms of direction and magnitude of the association of CHD with ART and twinning (eTable 3 in the Supplement).

    Discussion

    This large population-based study found that compared with infants born to mothers who conceived without using ART, infants born to mothers who conceived with ART were at higher risk of CHD. The association between ART and CHD was largely mediated by twinning. The contributions to overall excess risk from other associations of ART (not involving twinning) and from the ART-twin interaction were smaller and not statistically significant. However, the results do not rule out possible small contributions of those mechanisms to excess risk.

    Our mediation analysis is different from previous studies of the contribution of twinning in the literature. Previous studies40-43 have used different strategies to analyze the association of ART, twinning, and CHD, with most considering twinning as a confounder. Treating a mediator as a confounder in an analysis is flawed.40-43 Mediation analysis helped to assess the association between ART (exposure) and CHD (outcome). We believe that ignoring twinning as a mediator would lead to unexplainable or paradoxical findings. The finding that twinning contributed up to 87% of the association between ART and CHD may be important in the clinical practice of perinatal medicine. For example, instead of performing fetal echocardiography in ART pregnancies,32 it may be more cost-efficient to perform this procedure in twin pregnancies. Policy makers may consider a policy of single embryo transfer. Since 2015, Ontario government-provided public funding for IVF treatment required single embryo transfer in most women younger than 38 years.48 Some other adverse perinatal outcomes may share mechanisms of CHD,49 and a similar analytic strategy could be applied in future studies to assess these associations as well.

    Strengths and Limitations

    Our study has several strengths. First, the study sample was based on the entire population from 1 Canadian province (Ontario). Because BORN has 100% capture for hospital and home births in Ontario, our study was free from bias caused by selective inclusion of births. Second, the quality of BORN data has been evaluated by a number of studies,50 and it has been considered to be high. Third, the sample size of our study was the largest reported to date in this field. Fourth, we used multiple sources to identify CHD, including birth records, pregnancies that were lost or terminated after prenatal screening and in antenatal clinics (a total of 198 CHD cases were ascertained from losses or terminations), and cases diagnosed after birth (≤1 year) through record linkage with CIHI databases, with better ascertainment of outcome. Ascertainment of CHD cases in aborted fetuses might be more complete in ART pregnancies than in non-ART pregnancies, resulting in an artificially higher prevalence of CHD after ART. To address this possibility, we performed a sensitivity analysis restricted to pregnancy outcomes. The result of this sensitivity analysis was similar to the main results (eTable 3 in the Supplement), suggesting that ascertainment bias is not a serious threat to the validity of our study results. Fifth, rich demographic and clinical information collected by BORN allowed a thorough adjustment for confounding, assessment of mediation, and subgroup analyses. The use of mediation analysis helped to sort the associations of ART with twinning and CHD. Although the data contain no information on infertility, some important risk factors that are associated with both infertility and CHD have been adjusted. As a result, we were able to control some of the confounding by infertility, although not completely.

    Several limitations of our study should be recognized. First, some important variables, such as ethnicity, household income, and maternal education, were either not available or not measured at an individual level. Second, the higher socioeconomic status and healthy behaviors of couples who underwent ART compared with couples who conceived without ART may confound the observed results. Although we adjusted several maternal characteristics to mitigate confounding, administrative data cannot guarantee complete and accurate data collection. As a result, residual confounding may still exist. As VanderWeele and Ding50 suggested, a central concern about observational data are unmeasured or uncontrolled confounding, and with observational data, we can never be certain that the adjustments are adequate.51 Third, we were not able to address the issue of donor gametes, which may have a different association with the outcome for ART pregnancies compared with own gametes.1 Fourth, there were missing values for variables included in the regression analysis. We used multiple imputation to create a complete data set to mitigate the consequences of missing values in the analysis. However, analysis deleting records with missing variables or deleting CHDs ascertained from fetal losses or termination yielded similar results, suggesting the robustness of the data used in this study. Fifth, if the ART-twin interaction was present, both the natural indirect association and the natural direct association depended on the population frequency of twinning.43 The ART-twin interaction was small and not clinically significant, whereas the difference between the natural direct association and the natural indirect association was large and clinically significant, and the interpretation of the natural direct association and the natural indirect association was unaffected by the ART-twin interaction. Sixth, we were not able to address zygosity and chronicity of twins. Because the main focus of this study was the indirect association of twinning with CHD and ART, and zygosity and chronicity of twins may be a lesser concern in the assessment of indirect association than an assessment of direct association. Seventh, because CARTR Plus data on IVF cycles is not completely captured by BORN, we used information from BORN to supplement the ascertainment of ART. As a result, some ICSI cases in BORN may have been misclassified as IVF because BORN data sometimes lack details to differentiate ICSI from IVF. Because the focus of this study was an ART and non-ART comparison, it may be acceptable to have some IVF or ICSI misclassification to ensure a complete ascertainment of ART.

    Conclusions

    This study found that the observed association between ART and CHD may be substantially mediated by twinning. We believe that this finding is important to the clinical practice of perinatal medicine and to policy making. The robustness of this analytic strategy should be tested in the assessment of the association of ART with CHD in other populations and ART with other perinatal outcomes that may share similar mechanisms of CHD.

    Back to top
    Article Information

    Accepted for Publication: November 1, 2019.

    Published Online: February 24, 2020. doi:10.1001/jamapediatrics.2019.6096

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 Wen SW et al. JAMA Pediatrics.

    Corresponding Author: Shi Wu Wen, PhD, OMNI Research Group, Ottawa Hospital Research Institute, 501 Smyth Rd, Ottawa, ON K1H 8L6, Canada (swwen@ohri.ca).

    Author Contributions: Dr Miao had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Wen, Lougheed, Gaudet, Sprague, Walker.

    Acquisition, analysis, or interpretation of data: Wen, Miao, Taljaard, Davies, Lanes, Leader, Corsi.

    Drafting of the manuscript: Wen, Leader.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Miao, Taljaard.

    Obtained funding: Wen, Walker.

    Administrative, technical, or material support: Miao, Gaudet, Davies, Lanes, Leader, Sprague, Walker.

    Supervision: Wen, Lougheed.

    Conflict of Interest Disclosures: Dr Wen reported receiving grants from the Heart and Stroke Foundation of Canada and from the Canadian Institutes for Health Research (CIHR) during the conduct of the study. Dr Miao reported receiving grants from the Heart and Stroke Foundation of Canada and from CIHR during the conduct of the study. Dr Lanes reported receiving a salary paid in part by the CIHR grant from Ottawa Hospital Research Institute during the conduct of the study. Dr Sprague reported receiving grants from the Heart and Stroke Foundation of Canada and from CIHR during the conduct of the study. Dr Walker reported receiving grants from CIHR during the conduct of the study. No other disclosures were reported.

    Funding/Support: This study was funded by the Heart and Stroke Foundation of Canada (Dr Wen), with supplemental funding through grant FDN-148438 (Drs Wen and Walker) from the CIHR. The statistical analysis of the study was performed by Dr Miao and was paid in part by grant FDN-14838 from the CIHR.

    Role of Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    Disclaimer: Parts of this material are based on data and information compiled and provided by the Canadian Institute for Health Information. However, the analyses, conclusions, opinions and statements expressed herein are those of the author(s) and not necessarily this organization.

    Additional Contributions: Ruth Rennicks White, BNS, Natalie Rybak, BA, Alysha Harvey, BA, MSc, and Yanfang Guo, PhD (OMNI Research Group, Ottawa Hospital Research Institute, Ottawa, Ontario, Canada), helped in data analysis. Catherine Riddell, BA (BORN Ontario), helped in editing. None of these individuals were compensated for their help.

    References
    1.
    Allen  VM, Wilson  RD, Cheung  A; Genetics Committee; Reproductive Endocrinology and Infertility Committee.  Pregnancy outcomes after assisted reproductive technology.   J Obstet Gynaecol Can. 2006;28(3):220-233. doi:10.1016/S1701-2163(16)32112-0 PubMedGoogle ScholarCrossref
    2.
    Toner  JP, Coddington  CC, Doody  K,  et al.  Society for Assisted Reproductive Technology and assisted reproductive technology in the United States: a 2016 update.   Fertil Steril. 2016;106(3):541-546. doi:10.1016/j.fertnstert.2016.05.026 PubMedGoogle ScholarCrossref
    3.
    Bahtiyar  MO, Campbell  K, Dulay  AT,  et al.  Is the rate of congenital heart defects detected by fetal echocardiography among pregnancies conceived by in vitro fertilization really increased? a case-historical control study.   J Ultrasound Med. 2010;29(6):917-922. doi:10.7863/jum.2010.29.6.917 PubMedGoogle ScholarCrossref
    4.
    Giorgione  V, Parazzini  F, Fesslova  V,  et al.  Congenital heart defects in IVF/ICSI pregnancy: systematic review and meta-analysis.   Ultrasound Obstet Gynecol. 2018;51(1):33-42. doi:10.1002/uog.18932 PubMedGoogle ScholarCrossref
    5.
    Hansen  M, Kurinczuk  JJ, Bower  C, Webb  S.  The risk of major birth defects after intracytoplasmic sperm injection and in vitro fertilization.   N Engl J Med. 2002;346(10):725-730. doi:10.1056/NEJMoa010035 PubMedGoogle ScholarCrossref
    6.
    Hansen  M, Bower  C, Milne  E, de Klerk  N, Kurinczuk  JJ.  Assisted reproductive technologies and the risk of birth defects—a systematic review.   Hum Reprod. 2005;20(2):328-338. doi:10.1093/humrep/deh593 PubMedGoogle ScholarCrossref
    7.
    Hansen  M, Kurinczuk  JJ, Milne  E, de Klerk  N, Bower  C.  Assisted reproductive technology and birth defects: a systematic review and meta-analysis.   Hum Reprod Update. 2013;19(4):330-353. doi:10.1093/humupd/dmt006 PubMedGoogle ScholarCrossref
    8.
    Lancaster  PA.  Congenital malformations after in-vitro fertilisation.   Lancet. 1987;2(8572):1392-1393. doi:10.1016/S0140-6736(87)91277-3 PubMedGoogle ScholarCrossref
    9.
    Wen  SW, Leader  A, White  RR,  et al.  A comprehensive assessment of outcomes in pregnancies conceived by in vitro fertilization/intracytoplasmic sperm injection.   Eur J Obstet Gynecol Reprod Biol. 2010;150(2):160-165. doi:10.1016/j.ejogrb.2010.02.028 PubMedGoogle ScholarCrossref
    10.
    Panagiotopoulou  O, Fouzas  S, Sinopidis  X, Mantagos  SP, Dimitriou  G, Karatza  AA.  Congenital heart disease in twins: the contribution of type of conception and chorionicity.   Int J Cardiol. 2016;218:144-149. doi:10.1016/j.ijcard.2016.05.029 PubMedGoogle ScholarCrossref
    11.
    Votava-Smith  JK, Glickstein  JS, Simpson  LL, Williams  IA.  Comparison of method of conception in fetuses undergoing echocardiography at a tertiary referral center.   Prenat Diagn. 2014;34(5):445-449. doi:10.1002/pd.4327 PubMedGoogle ScholarCrossref
    12.
    Tararbit  K, Houyel  L, Bonnet  D,  et al.  Risk of congenital heart defects associated with assisted reproductive technologies: a population-based evaluation.   Eur Heart J. 2011;32(4):500-508. doi:10.1093/eurheartj/ehq440 PubMedGoogle ScholarCrossref
    13.
    Tararbit  K, Lelong  N, Thieulin  AC,  et al; EPICARD Study Group.  The risk for four specific congenital heart defects associated with assisted reproductive techniques: a population-based evaluation.   Hum Reprod. 2013;28(2):367-374. doi:10.1093/humrep/des400 PubMedGoogle ScholarCrossref
    14.
    Tararbit  K, Lelong  N, Houyel  L, Bonnet  D, Goffinet  F, Khoshnood  B; EPICARD Study Group.  Assessing the role of multiple pregnancies in the association between tetralogy of Fallot and assisted reproductive techniques: a path-analysis approach.   Orphanet J Rare Dis. 2014;9:27. doi:10.1186/1750-1172-9-27 PubMedGoogle ScholarCrossref
    15.
    Davies  MJ, Moore  VM, Willson  KJ,  et al.  Reproductive technologies and the risk of birth defects.   N Engl J Med. 2012;366(19):1803-1813. doi:10.1056/NEJMoa1008095 PubMedGoogle ScholarCrossref
    16.
    Simeone  RM, Tinker  SC, Gilboa  SM,  et al; National Birth Defects Prevention Study.  Proportion of selected congenital heart defects attributable to recognized risk factors.   Ann Epidemiol. 2016;26(12):838-845. doi:10.1016/j.annepidem.2016.10.003 PubMedGoogle ScholarCrossref
    17.
    Cedergren  MI, Källén  BA.  Maternal obesity and infant heart defects.   Obes Res. 2003;11(9):1065-1071. doi:10.1038/oby.2003.146 PubMedGoogle ScholarCrossref
    18.
    Pasquali  R, Patton  L, Gambineri  A.  Obesity and infertility.   Curr Opin Endocrinol Diabetes Obes. 2007;14(6):482-487. doi:10.1097/MED.0b013e3282f1d6cb PubMedGoogle ScholarCrossref
    19.
    Hur  YM, Song  TB.  A recent rise in twin birth rates and demographic changes in mothers of twins in South Korea: 2003-2007.   Twin Res Hum Genet. 2009;12(1):118-122. doi:10.1375/twin.12.1.118 PubMedGoogle ScholarCrossref
    20.
    Kurosawa  K, Masuno  M, Kuroki  Y.  Trends in occurrence of twin births in Japan.   Am J Med Genet A. 2012;158A(1):75-77. doi:10.1002/ajmg.a.34362 PubMedGoogle ScholarCrossref
    21.
    Dawson  AL, Tinker  SC, Jamieson  DJ,  et al; National Birth Defects Prevention Study.  Twinning and major birth defects, National Birth Defects Prevention Study, 1997-2007.   J Epidemiol Community Health. 2016;70(11):1114-1121. doi:10.1136/jech-2015-206302 PubMedGoogle ScholarCrossref
    22.
    Huang  CT, Au  HK, Chien  LW, Chang  CW, Chien  YY, Tzeng  CR.  Twin pregnancy outcome among cases of spontaneous conception, intrauterine insemination, and in vitro fertilization/intracytoplasmic sperm injection.   Fertil Steril. 2006;86(4):1017-1019. doi:10.1016/j.fertnstert.2006.02.109 PubMedGoogle ScholarCrossref
    23.
    Thurin  A, Hausken  J, Hillensjö  T,  et al.  Elective single-embryo transfer versus double-embryo transfer in in vitro fertilization.   N Engl J Med. 2004;351(23):2392-2402. doi:10.1056/NEJMoa041032 PubMedGoogle ScholarCrossref
    24.
    Koudstaal  J, Bruinse  HW, Helmerhorst  FM, Vermeiden  JP, Willemsen  WN, Visser  GH.  Obstetric outcome of twin pregnancies after in-vitro fertilization: a matched control study in four Dutch university hospitals.   Hum Reprod. 2000;15(4):935-940. doi:10.1093/humrep/15.4.935 PubMedGoogle ScholarCrossref
    25.
    Pharoah  PO.  Twins and cerebral palsy.   Acta Paediatr Suppl. 2001;90(436):6-10. doi:10.1111/j.1651-2227.2001.tb01620.x PubMedGoogle ScholarCrossref
    26.
    Hardin  J, Carmichael  SL, Selvin  S, Lammer  EJ, Shaw  GM.  Increased prevalence of cardiovascular defects among 56,709 California twin pairs.   Am J Med Genet A. 2009;149A(5):877-886. doi:10.1002/ajmg.a.32745 PubMedGoogle ScholarCrossref
    27.
    Bernier  PL, Stefanescu  A, Samoukovic  G, Tchervenkov  CI.  The challenge of congenital heart disease worldwide: epidemiologic and demographic facts.   Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu. 2010;13(1):26-34. doi:10.1053/j.pcsu.2010.02.005 PubMedGoogle ScholarCrossref
    28.
    Canobbio  MM.  Health care issues facing adolescents with congenital heart disease.   J Pediatr Nurs. 2001;16(5):363-370. doi:10.1053/jpdn.2001.26570 PubMedGoogle ScholarCrossref
    29.
    Lui  GK, Rogers  IS, Ding  VY,  et al.  Risk estimates for atherosclerotic cardiovascular disease in adults with congenital heart disease.   Am J Cardiol. 2017;119(1):112-118. doi:10.1016/j.amjcard.2016.09.023 PubMedGoogle ScholarCrossref
    30.
    Jay  PY, Akhirome  E, Magnan  RA,  et al.  Transgenerational cardiology: one way to a baby’s heart is through the mother.   Mol Cell Endocrinol. 2016;435:94-102. doi:10.1016/j.mce.2016.08.029 PubMedGoogle ScholarCrossref
    31.
    Kloesel  B, DiNardo  JA, Body  SC.  Cardiac embryology and molecular mechanisms of congenital heart disease: a primer for anesthesiologists.   Anesth Analg. 2016;123(3):551-569. doi:10.1213/ANE.0000000000001451 PubMedGoogle ScholarCrossref
    32.
    American Institute of Ultrasound in Medicine.  AIUM practice parameter for the performance of fetal echocardiography.   J Ultrasound Med. 2020;39(1):E5-E16.PubMedGoogle Scholar
    33.
    Better Outcomes Registry & Network (BORN). Ottawa, Ontario, Canada. http://www.bornontario.ca. Accessed March 25, 2019.
    34.
    Centers for Disease Control and Prevention, US Department of Health & Human Services. Assisted reproductive technology (ART). https://www.cdc.gov/art/whatis.html. Accessed March 25, 2019.
    35.
    Zhai  D, Guo  Y, Smith  G, Krewski  D, Walker  M, Wen  SW.  Maternal exposure to moderate ambient carbon monoxide is associated with decreased risk of preeclampsia.   Am J Obstet Gynecol. 2012;207(1):57.e1-57.e9. doi:10.1016/j.ajog.2012.03.022 PubMedGoogle ScholarCrossref
    36.
    Wilkins  R, Khan  S. Automated geographic coding based on the statistics Canada postal code conversion files: including postal codes through October 2010. https://mdl.library.utoronto.ca/sites/default/files/mdldata/open/canada/national/statcan/postalcodes/pccfplus/2006/2010oct/MSWORD.PCCF5H.pdf. Accessed July 15, 2019.
    37.
    Hernán  MA, Hernández-Díaz  S, Werler  MM, Mitchell  AA.  Causal knowledge as a prerequisite for confounding evaluation: an application to birth defects epidemiology.   Am J Epidemiol. 2002;155(2):176-184. doi:10.1093/aje/155.2.176 PubMedGoogle ScholarCrossref
    38.
    Schisterman  EF, Cole  SR, Platt  RW.  Overadjustment bias and unnecessary adjustment in epidemiologic studies.   Epidemiology. 2009;20(4):488-495. doi:10.1097/EDE.0b013e3181a819a1 PubMedGoogle ScholarCrossref
    39.
    Rothman  KJ, Greenland  S, Lash  TL.  Modern Epidemiology. 3rd ed. New York, NY: Wolters Kluwer/Lippincott Williams & Wilkins; 2008.
    40.
    VanderWeele  TJ.  Explanation in Causal Inference: Methods for Mediation and Interaction. New York, NY: Oxford University Press, 2015.
    41.
    VanderWeele  TJ.  Mediation analysis: a practitioner’s guide.   Annu Rev Public Health. 2016;37:17-32. doi:10.1146/annurev-publhealth-032315-021402 PubMedGoogle ScholarCrossref
    42.
    VanderWeele  TJ.  A unification of mediation and interaction: a 4-way decomposition.   Epidemiology. 2014;25(5):749-761. doi:10.1097/EDE.0000000000000121 PubMedGoogle ScholarCrossref
    43.
    Richiardi  L, Bellocco  R, Zugna  D.  Mediation analysis in epidemiology: methods, interpretation and bias.   Int J Epidemiol. 2013;42(5):1511-1519. doi:10.1093/ije/dyt127 PubMedGoogle ScholarCrossref
    44.
    van Buuren  S.  Multiple imputation of discrete and continuous data by fully conditional specification.   Stat Methods Med Res. 2007;16(3):219-242. doi:10.1177/0962280206074463 PubMedGoogle ScholarCrossref
    45.
    Little  RJA, Rubin  DB.  Statistical Analysis With Missing Data. 2nd ed. New York, NY: John Wiley & Sons; 2002. doi:10.1002/9781119013563
    46.
    Rubin  DB.  Multiple Imputation for Nonresponse in Surveys. New York, NY: John Wiley & Sons Inc; 2004.
    47.
    Muzyka  D, Lypka  M. The benefits of keeping it casual: investigating stress mediated health outcomes with PROC CAUSALMED. https://www.sas.com/content/dam/SAS/support/en/sas-global-forum-proceedings/2019/3897-2019.pdf. Accessed July 15, 2019.
    48.
    Procrea Fertility. Ontario treatment coverage and payment options. https://procrea.ca/financial-guidance/ontario-treatment-coverage-payment-options/ Accessed September 3, 2019.
    49.
    Oberg  AS, VanderWeele  TJ, Almqvist  C, Hernandez-Diaz  S.  Pregnancy complications following fertility treatment-disentangling the role of multiple gestation.   Int J Epidemiol. 2018;47(4):1333-1342. doi:10.1093/ije/dyy103 PubMedGoogle ScholarCrossref
    50.
    VanderWeele  TJ, Ding  P.  Sensitivity analysis in observational research: introducing the E-value.   Ann Intern Med. 2017;167(4):268-274. doi:10.7326/M16-2607 PubMedGoogle ScholarCrossref
    51.
    Better Outcomes Registry & Network. BORN data quality report, 2012-2014—executive summary. https://www.bornontario.ca/en/data/resources/Documents/BORN-Data-Quality-Report-2012-2014---Executive-Summary.pdf. Published April 2016. Accessed March 25, 2019.
    ×