Child health outcomes following fertility treatment and patient characteristics: discrepancies, limitations and prudence

Abstract

We argue that increased risk for a variety of diseases in ART children has been consistently reaffirmed by different methods and in diverse populations, providing a methodological critique of recent sibling studies, which hold great potential for studying the risks of ART. A recent within-family analysis using a national population register holds a distinct size advantage over previous studies, and suggested that apparent risks to offspring may be attributable to birth order. However, limitations of the design may have resulted in an erroneous conclusion. We discuss the advantages of a frequently neglected sibling study design, which compares siblings born of surrogate motherhood. While uncertainty remains, the evidence points to elevated risk for ART offspring. It may therefore be prudent to call for an extension of preventive and precautionary decisions to the entire population, and to change informed consent to incorporate the long-term health consequences of fertility treatments.

Keywords:

• National population register
• assisted human reproduction
• surrogate motherhood
• offspring health
• sibling studies
• informed consent

For more than one decade, the increased risk of diseases in children born from assisted reproductive technologies (ART) is progressively being consolidated in the scientific literature^1,2, with examples including Beckwith–Wiedemann Syndrome³, childhood cancer⁴, metabolic dysfunction⁵ and nonchromosomal birth defects⁶.

It was initially posited that the heightened risk of some diseases might be attributable to the incidence of multiple births and to the elevated age of women undergoing ART. But when additional studies^4–6 adjusted for these and other variables (race and ethnicity, parental educational level, child sex, paternal age, duration of breastfeeding, maternal and child body mass index, smoking during pregnancy, and family history of diabetes, stroke and heart disease, among others), the increase in risk remained. An alternative explanation attributed the increased risk of disease to confounding by characteristics of people struggling to conceive. That is, the causes of infertility, that led patients to seek treatment at ART clinics, were also responsible for the health problems of children born through the techniques.

Sibling studies may be particularly informative. Recently, it has been suggested that reduced birthweight in ART siblings could be explained by birth order⁷. This study is a within-family analysis using a national population register and, while the study holds a distinct size advantage over previous research, we believe the following limitations may have resulted in an erroneous conclusion.

First, the authors used an algorithm which has been validated for IVF but not for other types of fertility treatment⁸. Accordingly, the algorithm may misclassify participants, such that comparisons between siblings may not reflect desired comparisons between ART and natural conception (NC) where the only difference is mode of conception. For example, when only hormonal fertility treatment has been used, it is difficult to distinguish from the database whether the child was spontaneously conceived or through more intensive ART. The method used to establish ART exposure in the algorithm, which is to require a time difference of 44 weeks between the last cycle in which the drug was purchased and the birth of the child, does not let us know if the child was conceived by treatment or not. That is: some ART–NC siblings considered may be NC–NC siblings. While some misclassification is unavoidable, the degree of error in this case is not known, and could be considerable, causing the effects of ART to be underestimated. In addition, hormonal fertility treatment and intrauterine insemination (IUI) have less impact on child outcomes than IVF and intracytoplasmic sperm injection (ICSI), and a failure to make any distinction in this regard will dilute the effect of IVF/ICSI vs NC.

We also note that the sibling-pairs design does not control for time-varying confounding⁹. For example, the nature of ART treatment is likely to differ for primary compared to secondary infertility. Where ART has been used for a second birth, treatment protocols may vary compared to ART for patients who have not been able to conceive naturally. This might explain, to an unknown degree, the “order effect” observed by the authors. It is therefore crucial both that the finding of Goisis and colleagues⁷ is recognized as discrepant, and that possible explanations for the result are carefully considered.

Finally, we would highlight another relevant study, which is frequently neglected in discussions of sibling studies, but nonetheless falls into that category, which compared ART and NC siblings born of surrogate motherhood¹⁰. Distinct advantages of this study over other sibling studies include the fact that it identifies an effect in a fertile population (ruling out infertility confounding) and has no measurement error for exposure. This study showed that children born from commissioned embryos and carried by gestational surrogates have increased adverse perinatal outcomes (including preterm birth, low birth weight, hypertension, maternal gestational diabetes and placenta previa), compared with singletons conceived spontaneously and carried by the same woman, suggesting that assisted reproductive procedures may potentially affect embryo quality and that its negative impact cannot be overcome even with a proven healthy uterine environment. The study didn’t find statistical differences in the worst perinatal outcomes in those children coming from donor eggs and autologous eggs, suggesting that genetics don’t explain differences between ART and NC siblings.

While some uncertainty remains, the increased risk of several diseases in ART children (including, for example, Beckwith-Wiedemann Syndrome³, childhood cancer⁴, metabolic dysfunction⁵ and nonchromosomal birth defects [like fistula, atresia or stenosis]⁶) has been consistently reaffirmed by different methods and in diverse populations. Little by little, each link on the causal chain between manipulation at preconception time – when the epigenome is most vulnerable – and child diseases is being uncovered^11,12.

The scientific community has warned of the need to inform patients about the association between ART and offspring health for more than a decade¹³ and, as Barnhart pointed out, the responsibilities of ART clinics cannot be shifted onto the patient¹⁴. Calls to extend preventive and precautionary decisions to the entire population¹⁵ and to change informed consent to incorporate the long-term health consequences of fertility treatments are, in fact, increasingly frequent¹⁶ even in relation to techniques of low complexity, such as IUI¹⁷.

Transparency

Declaration of funding

There is no sponsorship/funding to disclose.

Declaration of financial/other relationships

No potential conflict of interest was reported by the authors. One peer reviewer has declared they are an employee of SPD Development Company Ltd., a fully owned subsidiary of SPD Swiss Precision Diagnostics GmbH, the manufacturer of home pregnancy and ovulation tests. CMRO peer reviewers on this manuscript have no other relevant financial or other relationships to disclose.

Author contributions

F.G.: literature research, study design, data analysis, data interpretation. J.W.: study design, data analysis, data interpretation and writing. J.H.: data analysis, data interpretation and final review. S.N.R.C.d.P.: literature research, study design and preparation for submission.

Acknowledgements

None stated.