Infertility and Risk of Congenital Anomalies: A Population-Based Cohort Study.

Milne and Shellenberger had full access to all of the data in the study and take responsibility for the integrity of the data and accuracy of the data analysis. Milne conducted the statistical analysis and drafted the work. The concept and design were developed by Milne, Brogly, Velez. Acquisition, analysis or interpretation of the data was performed by all authors. Velez obtained funding for the project. Brogly and Velez provided administrative, technical and material support, and reviewed the manuscript critically for important intellectual content.

The study was approved by the Queen's University Health Sciences and Affiliated Teaching Hospitals Ethics Research Board, Canada.

In this population‐based cohort study, infertility without treatment was associated with a small increased risk of any congenital anomaly, selected organ system anomalies, and specific anomalies. These findings add to the limited body of evidence suggesting that underlying infertility, independent of fertility treatment, may be a risk factor for congenital anomalies [ 8 ], and support the use of an appropriate infertility reference group when evaluating the safety of assisted reproductive technologies. The main strength of this study is its large sample size and population‐based design. Furthermore, our use of a clinically meaningful comparator group mitigated confounding by indication, a central challenge in studies evaluating the role of infertility treatments in offspring outcomes [ 3 ]. Births occurring before 20 weeks' gestation are not captured in BORN. If pregnancies among women with infertility were more likely than pregnancies from unassisted conceptions to end in spontaneous or therapeutic abortion due to congenital anomaly, selection bias may have occurred, resulting in underestimation of the true association. In addition, administrative health data do not capture the duration of attempts to conceive; therefore, the clinical definition of infertility (failure to conceive after 12 months of regular unprotected intercourse) could not be operationalized. Our definition of infertility has been used in previous studies [ 5 , 12 , 13 , 14 , 15 ], however, this approach may lead to exposure misclassification. Some women may not seek care or receive a diagnostic code for infertility, while others may consult earlier in the course of attempting conception. Furthermore, having health insurance has been found to be a predictor of seeking fertility treatment [ 40 ]; thus, excluding women without 5 years of OHIP coverage could introduce selection bias if women who did not meet our OHIP eligibility requirement did not have access to healthcare for such a diagnosis and were more likely to give birth to a baby with a congenital anomaly. These factors would likely attenuate associations toward the null. Finally, despite adjustment for baseline differences between women with and without infertility, residual confounding remains possible, as suggested by the E‐value analysis. Potential unmeasured confounders include race and paternal factors. Our results are consistent with those of US cohort studies that observed a small increased risk of any congenital anomaly among births to women with infertility who did not use fertility treatment compared to unassisted conceptions (adjusted odds ratio [OR] 1.15, 95% CI 1.08, 1.23; OR 1.14, 95% CI 1.06, 1.24) [ 4 , 7 ]. However, associations with any congenital anomaly are less informative because this broad classification may not capture differences in aetiology [ 17 ]. A cohort study from the Netherlands reported similar increased risks of specific anomalies, including abdominal wall anomalies (OR 2.43, 95% CI 1.05, 5.62), penoscrotal hypospadias (OR 9.83, 95% CI 3.58, 27.04), and right ventricular outflow obstruction (OR 1.77, 95% CI 1.06, 2.97) [ 41 ]. Additionally, a cohort study examining causes of infertility and the risk of congenital anomalies found that all infertility diagnoses were associated with increased risk (adjusted prevalence ratio [PR] 1.38, 95% CI 1.19, 1.59), with the highest risk in births to women with endometriosis (PR 1.81, 95% CI 1.36, 2.40) [ 42 ], similar to a recent study by our group [ 43 ]. Potential mechanisms may involve altered endometrial receptivity and early placental dysfunction in pregnancies affected by infertility, which could impair nutrient exchange and fetal organogenesis [ 44 ].

This study was funded by a grant MFM‐146444 from the Canadian Institutes of Health Research Institute of Human Development, Child & Youth Health, Clinician–Investigator Teams in Obstetrics and Maternal‐Fetal Medicine. Bailey Milne received funding for this project from the Queen's University Dean's Doctoral Award. This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health (MOH) and the Ministry of Long‐Term Care (MLTC). This document used data adapted from the Statistics Canada Postal Code OM Conversion File, which is based on data licensed from Canada Post Corporation, and/or data adapted from the Ontario Ministry of Health Postal Code Conversion File, which contains data copied under licence from Canada Post Corporation and Statistics Canada. Parts of this material are based on data and/or information compiled and provided by CIHI, the Ontario Ministry of Health, and the Better Outcomes Registry and Network (“BORN”), part of the Children's Hospital of Eastern Ontario. The analyses, conclusions, opinions, and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred.

This population‐based cohort was created using administrative health data from Ontario, Canada (Table  S1 ). Pregnancy characteristics, parity, and neonatal outcomes were obtained from the Better Outcomes Registry & Network Ontario (BORN), which captures 99% of hospital births and serves as the Ontario Ministry of Health's mandatory perinatal data registry [ 9 ]. All live births and stillbirths at ≥ 20 weeks' gestation to women aged 15–50 between 2006 and 2021 were included. Births conceived through fertility treatment were excluded because the objective of this study was to examine infertility independent of fertility treatment. Births to gestational carriers were also excluded [ 10 , 11 ]. Maternal demographics and pre‐existing health conditions were obtained by linking ambulatory, hospital, and surgical databases. Mother‐infant pairs with < 5 years of Ontario Health Insurance Program (OHIP) eligibility prior to conception were excluded to prevent under‐ascertainment of infertility diagnosis. Pregnancies with missing information on number of foetuses were excluded; however, those with missing stillbirth status ( N  = 139) remained in the cohort. Exposure was categorised as (i) unassisted conception, or (ii) infertility without treatment, defined as a history of one or more infertility consultations with a physician within 2 years before conception (OHIP physician billing code ICD‐9628), in the absence of any fertility treatment recorded in the BORN delivery record [ 5 , 12 , 13 , 14 , 15 ]. Congenital anomalies were identified from the Canadian Institute for Health Information Infant Discharge Abstract Database using ICD‐10 codes Q000‐Q999 and classified according to the Metropolitan Atlanta Congenital Defects Program algorithm [ 16 , 17 ] as follows: (i) any congenital anomaly, (ii) organ system anomalies, and (iii) specific anomalies. We adjusted for potential confounders, including maternal age at delivery, income quintile at delivery, rurality, pre‐existing diabetes, chronic hypertension, obesity prior to pregnancy, and smoking, alcohol use, and substance use during pregnancy (Figure  1 ). The rationale for including these covariates is provided in Table  S2 [ 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 ]. Conceptual diagram for association between subfertility and risk of congenital anomalies. Modified Poisson regression, using a robust error variance to account for multiple births to the same woman during the study period, was used to estimate adjusted relative risks (RRs, binomial link) and adjusted risk differences (RD, identity link) of anomalies in births to women with infertility who did not use treatment compared to births following unassisted conceptions. Given the increased risk of congenital anomalies among multifetal births, a sensitivity analysis was conducted restricting the cohort to singleton births. An additional sensitivity analysis excluded smoking and alcohol variables because these were measured during pregnancy and used as a proxy for preconception confounders. Finally, an E‐value analysis was performed to assess the potential influence of unmeasured confounding from race. All statistical analyses were performed using SAS (version 9.4; Cary, NC, USA). Missingness was 0% for all covariates included in the analysis (Table  S3 ). Stillbirth status was missing for 0.01% of pregnancies.

There were 1,415,324 births included in the cohort: 1,185,299 (83.7%) were from unassisted conceptions and 230,025 (16.3%) were to women with infertility who did not use fertility treatment (Figure  2 ). Compared with women who conceived without assistance, women with infertility were older, more likely to have a higher income, reside in an urban area, and have pre‐existing diabetes or chronic hypertension, and were less likely to smoke, consume alcohol, or other substances during pregnancy. Nulliparity was more common in unassisted conceptions (Table  1 ). The prevalence of congenital anomalies was 5.3% among births from unassisted conceptions and 5.9% among births to women with infertility. Cohort creation flowchart. Characteristics of infants born in Ontario, 2006–2021, according to exposure group. Reported as n (%), unless otherwise specified. Relative to births from unassisted conceptions, births to women with infertility who did not use fertility treatment had a small increased risk of congenital anomalies overall and across multiple organ systems. Small increased risks were observed for cardiovascular, central nervous system, eye, ear, face, neck, gastrointestinal, genital, musculoskeletal, urinary, and neoplastic anomalies. For specific anomalies, a small increase in risk was observed for atrial septal anomalies, ventricular septal anomalies, stenosis of pulmonary arteries, unspecified cleft palates, ankyloglossia, craniosynostosis, plagiocephaly, kidney malformations, undescended testes, other penis malformation, unspecified hypospadias, inguinal hernias, and pilonidal cysts (Table  2 ). After adjustment for covariates, the absolute risk difference for any congenital anomaly was 6.06 per 1000 births (95% CI 4.96, 7.14) (Table  S4 ). Adjustment for potential confounders increased some estimates relative to the unadjusted analyses, while some estimates were attenuated after adjustment (Table  2 and Table  S4 ). Relative risk of anomalies for infertile births without treatment compared to unassisted conception. Relative risks were adjusted for maternal age at delivery, income quintile at delivery, rurality, pre‐existing diabetes, chronic hypertension, obesity prior to pregnancy, and smoking, alcohol use, and substance use during pregnancy. The proportion of multiple births was 1.0% among births from unassisted conceptions and 2.3% among births to women with infertility who did not use fertility treatment. When the cohort was restricted to singleton births, associations persisted with few exceptions (Table  S5 ). Increased risks were no longer observed for unspecified cleft palate or plagiocephaly. Excluding smoking and alcohol variables did not materially alter the results; only plagiocephaly was attenuated (Table  S6 ). The E‐value analysis indicated that an unmeasured confounder, such as race, would need to have associations of at least 1.37 (lower 95% CI: 1.21) to 2.79 (lower 95% CI: 1.99) with both the exposure and outcome to fully explain the observed associations (Table  S7 ).

Fertility treatments have been associated with an increased risk of congenital anomalies [ 1 , 2 ]. However, most studies have compared births conceived through fertility treatment with unassisted conceptions; consequently, the reported estimates reflect the combined effects of fertility treatment and the underlying infertility. This comparison introduces potential confounding by indication, as the reproductive conditions necessitating fertility treatment may independently influence the risk of congenital anomalies [ 3 ]. Infertility is defined as the inability to conceive after 12 months of regular unprotected sexual intercourse [ 4 ]. Some women with infertility conceive without fertility treatment [ 5 , 6 ]. Recent evidence suggests that infants born to women with infertility who do not use fertility treatment may have an increased risk of congenital anomalies [ 4 , 7 ]. However, most studies have focused on any congenital anomaly, whereas associations with organ system and specific anomalies have been less frequently reported [ 8 ]. Therefore, we used a large population‐based cohort with universal healthcare coverage to estimate the association between infertility, independent of fertility treatment, and the risk of organ system and specific congenital anomalies compared with unassisted conception.

Although the overall risk of congenital anomalies was low, births to women with infertility who did not use fertility treatment had a small increased risk of any anomaly, selected organ system anomalies, and specific anomalies compared with births from unassisted conceptions. Future studies should consider specific infertility diagnoses, such as polyendocrine metabolic syndrome, endometriosis, or male factor infertility, on the risk of organ system and specific congenital anomalies.

The authors declare no conflicts of interest.

Table S1: List of Cohort Creation Variable, Study Exposure and Outcome Variables, and Model Covariates. Table S2: Covariates identified a priori as potential confounders. Table S3: Missingness of data on characteristics of the cohort. Table S4: Risk differences of congenital anomalies per 1000 births to infertile women without treatment compared to unassisted conceptions. Table S5: Relative risk of anomalies for births to infertile women without treatment compared to unassisted conception restricted to singleton births. Table S6: Relative risk of anomalies for births to infertile women without treatment compared to unassisted conception with smoking and alcohol covariates removed from the model. Table S7: E‐Value sensitivity analysis for risk of anomalies.