The Use of Assisted Reproductive Technology by European Childhood Cancer Survivors

CCS often wish to have biological children yet harbour concerns about fertility impairment, pregnancy risks and the general health risks of prospective offspring. To clarify these concerns, health outcomes in survivor offspring born following ART (n = 74, 4.5%) or after spontaneous conception (n = 1585) were assessed in our European offspring study by descriptive and bivariate analysis. Outcomes were compared to a sibling offspring cohort (n = 387) in a 4:1 matched-pair analysis (n = 1681). (i) Survivors were more likely to employ ART than their siblings (4.5% vs. 3.7%, p = 0.501). Successful pregnancies were achieved after a median of one cycle with, most commonly, intracytoplasmic sperm injection (ICSI) using non-cryopreserved oocytes/sperm. (ii) Multiple-sibling births (p < 0.001, 29.7% vs. 2.5%), low birth weight (p < 0.001; OR = 3.035, 95%-CI = 1.615–5.706), and preterm birth (p < 0.001; OR = 2.499, 95%-CI = 1.401–4.459) occurred significantly more often in survivor offspring following ART utilisation than in spontaneously conceived children. ART did not increase the prevalence of childhood cancer, congenital malformations or heart defects. (iii) These outcomes had similar prevalences in the sibling population. In our explorative study, we could not detect an influence on health outcomes when known confounders, such as multiple births, were taken into account.


Introduction
In recent decades, improved treatment has increased childhood cancer survival across Europe. This has led to a growing number of adults requiring specific life-event counselling, such as family planning. Many individuals of childbearing age with a history of cancer wish to have biological children. Still, they are faced with concerns regarding fertility impairment, potential pregnancy risks and the overall health outcome of their offspring [1,2]. Exposure to high-dose chemotherapy or gonadal irradiation can impair fertility or cause permanent infertility, depending on the type of agent, cumulative dosage, patient age and gender [3]. As a consequence, reduced overall pregnancy and live birth rates in former patients compared to the general population have been reported [4,5]. While some childhood cancer survivors conceive naturally, others require the use of assisted reproductive technology (ART), such as in vitro fertilisation (IVF) and/or intracytoplasmatic sperm injection (ICSI), which represent the most commonly used methods. Survivors who achieve pregnancy may experience a range of associated risks to both mother and foetus [6,7], including an increased risk of spontaneous abortion, stillbirth and preterm birth [8][9][10]. These findings were reported mainly in survivors treated with high-dose abdominal irradiation [11,12]. It is encouraging to note that no increased risk of congenital abnormalities was observed in these offspring [13,14]. To date, the adverse impact of cancer treatment on pregnancy outcomes has only been investigated in selected patient groups based on diagnosis or age at diagnosis; however, effects on the long-term health of survivor offspring remain unclear.
In 2019, a global estimation of more than 7 million ART-conceived children was documented; this number is steadily rising and comprises 2-6% of European births [15]. However, concerns exist regarding the health and development of ART-conceived children [16]. Multiple pregnancies resulting from multiple embryo transfers are associated with low birth weight and preterm birth, which can have long-term health implications [17,18]. However, ART has also been suspected of adverse obstetric and perinatal consequences of single births [19,20], such as increased risk of birth defects [21], cancer [22] and early growth [23]. Further evidence suggests that ART treatment also increases the risk of chronic age-related diseases such as obesity, type 2 diabetes and cardiovascular disease [24][25][26][27]. However, the influence of parental characteristics, including the aetiology of subfertility and infertility, or aspects of specific infertility treatments are unknown. The underlying mechanisms, including epigenetic changes that may occur during preimplantation and the development of uterine reprogramming, are currently under debate [28,29]. Publications examining ART and health issues of childhood cancer survivor offspring are rare [30]. Our study compared: (i) the prevalence of ART utilisation in survivors and their siblings, (ii) perinatal outcomes, malformations, heart defects and cancer in ART-conceived survivor offspring with those conceived spontaneously and (iii) the outcomes of survivor offspring vs. sibling offspring.

Study Design and Participants
The European Offspring Study, an explorative, retrospective cohort study, surveyed both adult childhood cancer survivors and their adult siblings on the health of their biological offspring between 2013 and 2016 in five European countries (Germany, Austria, Czech Republic, Poland and Switzerland). Participants who gave their informed consent were surveyed using a 46-item questionnaire [31], designed to address five health-related subareas (diseases, health-related behavior, health-related quality of life, healthcare utilisation and living conditions) as well as socio-demographic information. The study was approved by the local ethics committees of participating centres (lead vote Charité-Universitätsmedizin Berlin, EA2/237/05, EA2/103/11). The detailed concept and study methods, including recruitment strategies and participant characteristics, have been described previously [32].

Variables
Our analyses included parental reports on offspring gender, year of birth, gestational age (categorized as per World Health Organization (WHO)), birth weight (categorized as per WHO), mode of conception, multiple births, diagnoses of congenital malformations and heart defects (both categorized as per International Classification of Disease (ICD-10)) [33], diagnoses of cancers (categorized as per The International Classification of Childhood Cancer (ICCC-3)). Parental characteristics were assessed, including educational attainment (classified as per ISCED [34], country of origin and the family's migrant background (fulfilled if the offspring, the parents or grandparents were born in another country than the country of study conduction)) and maternal smoking/alcohol consumption during pregnancy. Core data of the survivor parent (including date of birth, date of cancer diagnosis, cancer diagnosis and treatment) were additionally collected from medical records by participating centers. Parental age at diagnosis was grouped as 0-4, 5-9, and 10 years or older, and the type of cancer was classified as leukaemia/lymphoma, brain tumours or extra-cranial solid tumours. We conducted telephone interviews with patients and siblings who had used ART to obtain fertility cycle data, including the type of infertility factor, type of ART, number of cycles conducted, use of fresh, cryopreserved or donor sperm/oocytes, pregnancy complications and maternal age at birth of offspring.

Statistical Methods
Statistical analyses were carried out with IBM SPSS Statistics software, version 27 (IBM SPSS Statistics, Chicago, IL, USA); in addition, matched-pair analysis was conducted using R software, version 4.1.2. (R Software Inc., San Francisco, CA, USA). Questionnaires lacking information on the child's gender, age or mode of conception were excluded from analyses. Spontaneously conceived survivor offspring (n = 1585) and ART-conceived offspring (n = 74) were compared. For the additional comparison of survivor and sibling offspring, data on offspring lacking information on matching criteria (offspring gender and age, multiple births) were excluded. In total, 1294 survivor offspring were compared to 387 sibling offspring.
Perinatal and health outcomes of survivor offspring born following ART vs. spontaneous conception were analysed using descriptive statistics. p-values were calculated using the two-sided Chi-squared test and Pearson's correlation for non-parametric variables, and Spearman's correlation for parametric variables (level of significance: <0.05). Interaction effects were examined by binary logistic regression, which estimated adjusted odds ratios (ORs) and 95% confidence intervals (Cls). Binary logistic regression assessed the intervariable dependencies of the confounder's gender, age at the time of survey, migration background, ART, preterm birth, multiple births, congenital malformations, heart defects and parental estimation of offspring health as well as parental educational attainment, smoking/alcohol consumption during pregnancy, parental age at diagnosis and type of cancer. With these independent variables, four-fold logistic regression was carried out, respectively, for the following dependent variables: preterm birth, low birth weight, congenital malformation, and congenital heart defects. Furthermore, perinatal and health outcomes of survivor offspring were compared to those of sibling offspring using a 4:1 matched-pair analysis in a case-control design. Binary logistic regression with cluster data was performed for the matched sample.

Participants and (i) Characteristics of ART Utilisation
Overall, data from 1659 children born to childhood cancer survivors were included in our analyses, of which 74 were born following ART (4.5%). In this unmatched dataset, out of the total of 405 sibling offspring, 15 (3.7%) were born following ART. ART-conceived survivor offspring were significantly younger at the time of the survey (p < 0.001), more likely to be born a twin (p < 0.001), and none of the survivor parents reported smoking during pregnancy (p = 0.013) ( Table 1). Paired analysis of survivor offspring matched to sibling offspring showed that survivors consumed significantly less alcohol during pregnancy (p = 0.002, Table 1). Overall, 45 survivors were successfully interviewed regarding 51/74 ART-conceived offspring (68.9%, Table 2). All of these survivors had received chemotherapy, and two had also undergone radiotherapy.  In both childhood cancer survivors and siblings, male factor infertility was the more frequent reason for ART usage. The majority of survivors and siblings underwent ICSI (66.1%, 39/59) and used fresh oocytes/sperm (80.7%, 46/57). IVF and ICSI were successful after one cycle in half of the couples; however, 9.1% (4/44 pregnancies in survivors) required ≥4 cycles, whereas none of the siblings needed ≥4 cycles (Table 2).

Perinatal Outcomes
ART-conceived survivor offspring were born significantly more often preterm (<37 weeks of gestation, p <0.001) and with a birth weight below 2500 g (low birth weight, p < 0.001) compared to spontaneously conceived survivor offspring (Table 3). Multivariable analyses revealed that the prevalence of low birth weight and preterm birth were not associated significantly with ART. Multiple birth (p < 0.001), smoking during pregnancy (p = 0.014), congenital malformations (p = 0.004), older age at survey (p = 0.001) and solid tumours (p = 0.004) were associated with preterm birth in our cohort. Low birth weight was associated with preterm birth (p < 0.001, Table 4).   Paired analysis of survivor offspring with sibling offspring, independent of the mode of conception, revealed no significant differences in the prevalence of preterm birth and low birth weight (Table 3).

Prevalence of Childhood Cancer
Although ten children born to survivors (0.6%) were diagnosed with cancer (including two retinoblastomas in children with hereditary predispositions), none of the affected children were ART-conceived.
In paired analyses, only one child from the sibling offspring group was diagnosed with leukaemia (0.3%, Table 5).    * Miss = missing data; CI, confidence interval; ICCC-3, International Classification of Childhood Cancer, third revision (https://seer.cancer.gov/iccc/, accessed on 9 August 2022); SD, standard deviation; IQR, interquartile range; ICD-10, International Statistical Classification of Diseases, 10th revision; OR, odds ratio. a Diagnosis of congenital malformations of the circulatory system are reported in detail in congenital heart defects. b Children with multiple diagnoses appear more than once in the table. c Numbers (n) were too small to report meaningful percentages.

Prevalence of Congenital Malformations and Heart Defects
Neither congenital malformations nor heart defects were more prevalent in ARTconceived survivor offspring than spontaneously conceived offspring (Table 5). Congenital malformations were shown to be associated with the male gender (p = 0.028, Table 4).
There was no difference in the prevalence of malformations or congenital heart defects in the survivor vs. sibling offspring population (Table 5).

Discussion
Our study examined outcomes of ART use in childhood cancer survivors and their siblings as well as perinatal and health outcomes in their offspring. Survivors of childhood cancer are at increased risk for fertility impairment, and thus, numbers of ART utilisation in this cohort have increased in recent years. In a previous study, ART was used by twice as many survivors as in the general population [30]. Our current analyses revealed that childhood cancer survivors also requested ART more often (4.5%) than their siblings (3.7%). Male factor infertility was stated as the main reason for ART by the majority of survivors; this is in line with data from the general population [35].
ART techniques continue to improve, and rising overall success rates are beneficial to former cancer patients. Currently, one-third of pregnancies are successful after one cycle, one-half after two cycles and two-thirds after four. In our study, the survivors and their siblings were comparably successful. It is reassuring to note that pregnancies were achieved after only one cycle in half of our cases, especially as ART poses an additional psychological and physical burden to both former patients and their partners. The relevance of this is emphasized by the fact that the most commonly used method in our survivors was ICSI using fresh oocytes/sperm. This requires female hormonal stimulation, which can lead to 0.5 to 5% severe ovarian hyperstimulation syndrome [36]. However, it should be noted that in 16.9% of our cases, intrauterine insemination (IUI) was successful. Although this procedure also requires the preparation of both parental partners, it is less invasive. Furthermore, a recent study revealed no increased risks of congenital defects following IUI. Still, the underlying maternal infertility presented a potential elemental risk, in addition to the risk associated with IVF [20]. Not all of our patients succeeded after only a few cycles, and a proportion required 3 or more, which can also pose a financial burden. In Germany, statutory health insurers usually pay 50% of the costs for 3 cycles [37].
The overall success of human reproduction, whether spontaneous or after IVF, also depends strongly on the mother's age. In fact, for ART treatment, maternal age is one of the strongest predictors of success [38]. The main reasons for age-related infertility include reduced ovarian reserve and reduced oocyte/embryo competence due to age-related disorders, especially regarding an increased incidence of aneuploidy. Recent epidemiological studies confirmed that the adverse effects of chemotherapy on fertility are less severe when women seek pregnancy at a younger age [38]. In the collective studied, the average maternal age at first birth was below 30 years for both spontaneous conceptions and after ART.
To achieve higher pregnancy rates, the transfer of two or more embryos was previously the gold standard in ART. However, recent practices favour a single embryo transfer policy to avoid multiple births. The positive consequences of declining multiple birth rates after ART are decreasing perinatal risks such as preterm birth, intrauterine growth restriction and prenatal death, as well as decreasing risks for the mother, such as pre-eclampsia, diabetes and bleeding during labour [17]. Our study also showed the known increase in ART-conceived multiple births. When known confounders, including multiple births, were taken into account, perinatal outcomes, preterm births and low birth weight were no significantly different in survivor offspring, whether ART-conceived or spontaneous.
Smoking and drinking alcohol during pregnancy carries significant risks for the foetus, such as birth defects, premature birth, and low birth weight, and therefore should be avoided. We previously examined health behavior among childhood cancer survivors compared to the general population in a subgroup of this cohort. Parents who included a cancer survivor smoked less in the presence of their children. During pregnancy, mothers in cancer survivor parent couples abstained from drinking alcohol more often and smoked less [39]. Taking smoking and alcohol consumption into account as confounders, we also noticed that survivors were even more likely to abstain from smoking/alcohol during pregnancies following ART conception vs. spontaneous conception. Similarly, survivors consumed less alcohol during pregnancy than their siblings.
We did not see an increase in the prevalence of childhood cancer, congenital malformations or heart defects in ART-conceived survivor offspring compared to those conceived naturally. Similarly, we did not detect differences in survivor offspring compared to sibling offspring. A recent study also revealed no elevated risk of cancer in ART-conceived children [40].
The study setting among childhood cancer survivors in Europe posed certain limitations. Recruitment was mainly based on previous surveys which identified survivors with biological children, potentially causing a selection bias. A selection bias can generally lead to, among others, increased participation of higher-educated and female respondents in surveys. The latter was the case in our survey, as a non-responder analysis showed [32]. This approach was chosen to reduce the study burden for survivors. The questionnairebased setting could produce a recall bias that could reduce data accuracy. However, all survivors had been treated according to standardised trial protocols, for which treatment information was available. The number of siblings and participants using ART was low. However, we had a relatively large overall sample of offspring of survivors of childhood cancer. Some data were incomplete, particularly within the ART interview. Despite these limitations, our analyses offer new insights into health issues in offspring born to childhood cancer survivors, and the high response rate reflects the strong interest shown by survivors regarding these issues.

Conclusions
Our study shows encouraging results for survivors of childhood cancer that demonstrate that the vast majority of offspring born to survivors do not experience adverse perinatal outcomes or later health problems, independently of whether a conception was spontaneous or required ART. Our findings in the survivor cohort and in comparison to their siblings support that the use of ART by childhood cancer survivors does not put offspring at additional risk for adverse health outcomes, including childhood cancer, congenital malformations or heart defects. Against the backdrop of progressively reduced toxicity regimens, our findings appear particularly reassuring for patients treated with today's less toxic protocols for childhood cancer.
We saw an increased number of multiple births following ART to spontaneous conception in survivors. Current methods use fewer embryos per transfer to reduce these multiple birth rates, reducing associated adverse outcomes. The increasing numbers of childhood cancer survivors who turn to ART stress the importance of establishing an up-to-date information base for counselling childhood cancer patients and survivors, including this new and reassuring information.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.