Risk of Cancer in Children and Young Adults Conceived by Assisted Reproductive Technology

Mandy Spaan; Alexandra W. van den Belt-Dusebout; Marry M. van den Heuvel-Eibrink; Michael Hauptmann; Cornelis B. Lambalk; Curt W. Burger; Flora E. van Leeuwen; on behalf of the OMEGA-steering group


Hum Reprod. 2019;34(4):740-750. 

In This Article

Abstract and Introduction


Study Question: Do children conceived by ART have an increased risk of cancer?

Summary Answer: Overall, ART-conceived children do not appear to have an increased risk of cancer.

What is Known Already: Despite the increasing use of ART, i.e. IVF or ICSI worldwide, information about possible long-term health risks for children conceived by these techniques is scarce.

Study Design, Size, Duration: A nationwide historical cohort study with prospective follow-up (median 21 years), including all live-born offspring from women treated with subfertility treatments between 1980 and 2001.

Participants/Materials, Setting, Methods: All offspring of a nationwide cohort of subfertile women (OMEGA study) treated in one of the 12 Dutch IVF clinics or two fertility clinics. Of 47 690 live-born children, 24 269 were ART-conceived, 13 761 naturally conceived and 9660 were conceived naturally or through fertility drugs, but not by ART. Information on the conception method of each child and potential confounders were collected through the mothers' questionnaires and medical records. Cancer incidence was ascertained through linkage with The Netherlands Cancer Registry from 1 January 1989 until 1 November 2016. Cancer risk in ART-conceived children was compared with risks in naturally conceived children from subfertile women (hazard ratios [HRs]) and with the general population (standardized incidence ratios [SIRs]).

Main Results and the Role of Chance: The median follow-up was 21 years (interquartile range (IQR): 17–25) and was shorter in ART-conceived children (20 years, IQR: 17–23) compared with naturally conceived children (24 years, IQR: 20–30). In total, 231 cancers were observed. Overall cancer risk was not increased in ART-conceived children, neither compared with naturally conceived children from subfertile women (HR: 1.00, 95% CI 0.72–1.38) nor compared with the general population (SIR = 1.11, 95% CI: 0.90–1.36). From 18 years of age onwards, the HR of cancer in ART-conceived versus naturally conceived individuals was 1.25 (95% CI: 0.73–2.13). Slightly but non-significantly increased risks were observed in children conceived by ICSI or cryopreservation (HR = 1.52, 95% CI: 0.81–2.85; 1.80, 95% CI: 0.65–4.95, respectively). Risks of lymphoblastic leukemia (HR = 2.44, 95% CI: 0.81–7.37) and melanoma (HR = 1.86, 95% CI: 0.66–5.27) were non-significantly increased for ART-conceived compared with naturally conceived children.

Limitations, Reasons for Caution: Despite the large size and long follow-up of the cohort, the number of cancers was rather small for subgroup analyses as cancer in children and young adults is rare.

Wider Implications of the Findings: Overall, ART-conceived children do not appear to have an increased cancer risk after a median follow-up of 21 years. This large study provides important results, enabling physicians to better inform couples considering ART about the long-term safety of ART for their children. However, larger studies with prolonged follow-up are needed to investigate cancer risk in adults and in children conceived by ICSI and/or from cryopreserved embryos.

Study Funding/Competing Interest(s): This work was supported by The Dutch Cancer Society (NKI 2006–3631) which funded the OMEGA-women's cohort and Children Cancer Free (KIKA;147) which funded the OMEGA-offspring cohort. We declare no competing interests.


The use of ART, i.e. IVF and ICSI, has strongly increased in Western countries (Calhaz-Jorge et al., 2016; Sunderam et al., 2017). Over 6 million ART-conceived children have been born, including 80 000 in The Netherlands (Calhaz-Jorge et al., 2016; Sunderam et al., 2017).

Each phase of the ART procedure, including stimulation of multiple follicles, the process of oocyte retrieval and sperm preparation, culture of embryos, cryopreservation of sperm, oocytes and embryos, and embryo transfer, is substantially different from natural conception (Buitendijk, 1999). These processes occur in the same timeframe as epigenetic programming (Iliadou et al., 2011). During the first days after conception, the pre-implantation embryo is sensitive to environmental factors as created during embryo culture which may alter fetal development, with both prenatal and postnatal consequences (Geuns et al., 2007; Ceelen et al., 2008; Market-Velker et al., 2010). Additionally, differences in methylation of several genes have been found in cord blood and placenta samples from children conceived in vitro versus in vivo (Katari et al., 2009). Altered epigenetic patterns have been found in embryos derived from animals after ART(Farin et al., 2004; Iliadou et al., 2011). Furthermore, some studies of ART-conceived children reported an unexpectedly high incidence of rare cancer predisposition syndromes such as Beckwith–Wiedemann syndrome and Angelman syndrome (DeBaun et al., 2003; Gosden et al., 2003; Maher et al., 2003; Orstavik et al., 2003).

In 2013, a systematic meta-analysis showed an increased overall cancer risk in children born after fertility treatment (relative risk: 1.33; 95% CI: 1.08–1.63) (Hargreave et al., 2013b). However, this meta-analysis concluded that it is not clear whether other factors underlying subfertility rather than the ART procedure itself are the most important factors for (childhood) cancer. Therefore, more research was recommended (Hargreave et al., 2013b). Since then, several cohort studies reported reassuring results for overall cancer risk (Williams et al., 2013, 2018; Sundh et al., 2014; Hargreave et al., 2015; Reigstad et al., 2016; Lerner-Geva et al., 2017); however, two studies found an increased overall cancer risk among offspring of women with fertility problems (Hargreave et al., 2013a; Wainstock et al., 2017). While all studies found increased risks for specific cancer types (Williams et al., 2013; Hargreave et al., 2013a, 2015; Sundh et al., 2014; Reigstad et al., 2016; Lerner-Geva et al., 2017; Wainstock et al., 2017), these findings were inconsistent across studies. Studies had relatively short follow-up (median: 10 years [range: 6.6–18.0]) and other limitations such as a small number of cases (median: 119 [range: 7–481]) or restriction to a general population comparison group (no possibility for confounding adjustment) (Williams et al., 2013, 2018; Hargreave et al., 2013a, 2015; Sundh et al., 2014; Reigstad et al., 2016; Lerner-Geva et al., 2017; Wainstock et al., 2017), precluding firm conclusions (Reigstad et al., 2017). We therefore evaluated cancer risk in ART-conceived children and young adults in the Dutch nationwide OMEGA-offspring cohort (van den Belt-Dusebout et al., 2016). This cohort is uniquely positioned to examine this important question as it includes the oldest Dutch generation of ART-conceived children and a comparison group of naturally conceived children born to subfertile women.