The Developmental Origins of Health and Disease (DOHaD) theory hypothesizes that early-life exposures influence health and disease risk later in life, with epigenetic changes potentially playing a key role in fetal programming mechanisms [1]. Low birth weight and preterm birth, both indicative of impaired intrauterine growth, along with rapid postnatal catch-up growth, are strong predictors of adverse cardiometabolic risk factors and chronic diseases in adulthood [2–4]. Therefore, understanding childhood growth trajectories is essential for assessing development and health across the life course.

Several studies have suggested that children conceived through medically assisted reproduction (MAR) have a shorter gestational age, lower birth weight, and a higher risk of being small for gestational age compared to those conceived naturally [5]. It has been hypothesized that these differences may result from factors such as ovarian stimulation, culture media, specific MAR procedures (e.g., intracytoplasmic sperm injection [ICSI] or frozen embryo transfer [FET]) [6–8], as well as the underlying causes of subfertility or infertility [9]. However, a systematic review by Bay et al. on the long-term growth of singletons found no significant differences in weight or height between children conceived through MAR and those conceived naturally [10].

In twin pregnancies, a recent systematic review by Marleen et al. found that MAR is associated with a higher risk of maternal complications but has a varied impact on perinatal outcomes compared to natural conception [11]. Furthermore, the association between MAR and growth patterns in twins remains unclear, with conflicting findings in the literature. For example, a Chinese cohort study of 26,818 twins reported no significant differences in growth from birth to 18 years between MAR-conceived and naturally conceived twins [12]. In contrast, a British cohort study of 13,528 twins suggested that MAR twins born after FET grew faster than those born after fresh embryo transfer or natural conception [13].

Given these inconsistencies, our study aimed to compare growth trajectories from birth to 12 years between children conceived via infertility treatments and those conceived naturally. We further stratified the analysis by singletons and multiples. This study was conducted using data from a long-term, population-based national birth cohort in Taiwan.

We identified several key findings. First, although children conceived via IUI or IVF appeared to have lower weight, shorter height, and lower BMI than those conceived naturally when all children were analyzed together, these differences were primarily attributable to the higher proportion of multiple births in the IUI and IVF groups. Second, after stratifying by plurality, no significant differences in weight, height, or BMI were observed between the infertility treatment groups and the NC group. Among multiples, children in the IUI group might have slightly lower weight than those in the NC group. Third, the risks of rapid weight gain (6 months–2 years) and underweight, overweight, and obesity (5.5–12 years) did not significantly differ by infertility treatment status in either singletons or multiples. Overall, after accounting for plurality, we found no significant differences in growth trajectories between children conceived through infertility treatments and those conceived naturally.

The DOHaD hypothesis suggests that prenatal and perinatal exposure to environmental factors—such as undernutrition, stress, or environmental chemicals—can alter an individual’s phenotype and increase the risk of non-communicable diseases in adulthood [20]. Concerns have been raised that some MAR procedures may directly affect embryos, potentially impairing fetal and childhood growth. For example, IVF exposes gametes and embryos to various hormones, culture media, and in vitro manipulation techniques, including prolonged embryo culture, cryopreservation and thawing, prenatal genetic testing, and embryo biopsy [21]. Laboratory evidence indicates that these procedures can induce molecular changes, alter epigenetics and gene expression, and may influence placental and fetal development [22]. Therefore, monitoring the long-term health and growth of children conceived via infertility treatments is essential.

Most of our data showed that after accounting for plurality, infertility treatments were not significantly associated with childhood growth trajectories, suggesting that the higher rate of multiples, rather than the infertility treatments themselves, was the key factor related to impaired child growth. These findings demonstrate that the influence of MAR on long-term offspring growth is negligible if the rate of multiple births is controlled. Our findings also highlight the importance of minimizing multiple pregnancies in MAR. Singleton pregnancies result in better maternal and child outcomes than multiple pregnancies [23]. Therefore, most professional reproductive medicine societies, including those in Taiwan, advocate for elective single embryo transfer in IVF to reduce pregnancy risks associated with multiple births. As a result, the rate of multiples following IVF is declining globally [24,25].

In our analysis stratified by plurality, we found that multiples conceived via IUI had significantly lower weight than those conceived naturally. This finding aligns with research by Yeung et al., who observed that twins conceived via ovulation induction (with or without IUI) were smaller from birth to three years of age than naturally conceived twins [16]. Ovarian stimulation for IUI commonly involves clomiphene citrate, letrozole, or injectable gonadotropins, with gonadotropin use significantly increasing the risk of multiple gestation [26]. Ovarian hyperstimulation may lead to ovulation of immature oocytes or alter endometrial function due to exposure to high hormone levels (e.g., estradiol and progesterone), resulting in impaired placental function and restricted fetal growth [16]. Supporting this hypothesis, studies have shown that frozen embryo transfers—where endometrial preparation is more natural—carry a lower risk of small-for-gestational-age infants compared to fresh embryo transfers [27].

The major strengths of this study include its large, nationally representative cohort from Taiwan (n = 21,457), reducing selection bias. Participants were followed from 6 months of age, with repeated weight and height measurements enabling longitudinal analysis. Many previous studies lacked sufficient sample sizes or long-term follow-up, making our study a valuable addition to the literature. However, the numbers of children conceived via infertility treatment among singletons were relatively small. As a result, non-significant findings in singleton analyses should be interpreted with caution, as they do not necessarily indicate equivalence but may reflect limited statistical power and wide confidence intervals.

Study limitations should be considered. Infertility treatment status was self-reported rather than sourced from medical records, and child growth data were reported by caregivers rather than directly measured by healthcare professionals, introducing potential measurement errors, which might have biased the results towards the null. Additionally, children conceived by infertility treatments had shorter gestational ages and lower birth weights than naturally conceived children, even after stratifying by plurality (see Table 1). We did not adjust for these variables, as they are intermediate factors in the causal pathway. The observed lower weight in multiples conceived via IUI may partially reflect these differences.

Furthermore, although we adjusted for several sociodemographic and perinatal factors, information on maternal and paternal alcohol consumption and maternal nutrition during pregnancy was not available in the datasets accessed for this study and therefore could not be considered. In addition, breastfeeding was assessed only as a binary variable, as data on duration or exclusivity were not available. Finally, detailed information on the underlying etiology or pathophysiology of infertility among parents who underwent infertility treatment was not captured in the available data. These unmeasured factors may contribute to residual confounding and warrant further investigation in future studies with more detailed clinical and behavioral information.

Our findings contribute to the literature on long-term growth outcomes among children conceived through infertility treatments. Consistent with prior studies, we observed no clear evidence of differences in weight, height, or BMI after accounting for plurality, suggesting that infertility treatment itself is unlikely to have a clinically relevant impact on childhood growth. Instead, the higher prevalence of multiple births largely explains the differences observed in un-stratified analyses. This finding reassures couples undergoing infertility treatments. However, we lacked detailed MAR-related data, such as whether ICSI was used or whether embryo transfer was fresh or frozen. Future efforts should focus on establishing national MAR registries and birth cohort studies to assess the long-term effects of evolving MAR techniques on child growth and health outcomes. Importantly, our findings emphasize the need to minimize multiple births following MAR to optimize pregnancy and child health outcomes.