Fertility Treatments and the Risk of Kawasaki Disease Among Offspring

The present cohort study was based on data obtained from the Japan Environmental and Children’s Study (JECS), an ongoing nationwide prospective birth cohort study that recruited pregnant women between January 2011 and March 2014. Statistical analyses were performed between April and November 2025. Fifteen Regional Centres that covered northern to southern Japan (Hokkaido, Miyagi, Fukushima, Chiba, Kanagawa, Koshin, Toyama, Aichi, Kyoto, Osaka, Hyogo, Tottori, Kochi, Fukuoka, and South Kyushu and Okinawa) were selected. The JECS protocol and profile have been described elsewhere. 24 , 25 This study was conducted according to the guidelines in the Declaration of Helsinki. 26 The JECS protocol was reviewed and approved by the Ministry of the Environment’s Institutional Review Board on Epidemiological Studies and the ethics committees of all participating institutions. Written informed consent was obtained from all participants. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology ( STROBE ) reporting guideline. KD was included as one of the prespecified pediatric outcomes in the JECS questionnaires from the outset, with parental self-administered questionnaires conducted every 6 months until 6 years of age; questionnaires at ages 6, 12, 18, 24, 36, 48, and 60 months included items regarding KD history, and a secondary survey with the medical institutions listed in the questionnaire confirmed the diagnosis and acquired more detailed information. The cases confirmed through this secondary survey, and with onset up to 59 months of age were included in the present study. Among 1329 confirmed events of KD onset, we excluded recurrent events (n = 46) and those in which the first onset was out of the defined age range (n = 57), resulting in 1226 cases of first-onset KD. All conceptions were categorized as either spontaneous conception (SC) or assisted conception (AC), which included 4 types of fertility treatments: ovulation induction, intrauterine insemination, in vitro fertilization, and intracytoplasmic sperm injection. These 4 fertility treatment methods fall under the Medical Subject Headings classification of “Reproductive Techniques, Assisted.” Throughout this manuscript, we use the term AC or fertility treatments to refer collectively to all 4 methods. Information on fertility treatments was identified through medical record transcripts and self-administered questionnaires filled out by mothers during pregnancy. The former items were transcribed by physicians, nurses, midwives, hospital staff, or research coordinators involved in the JECS. When discrepancies occurred, more advanced treatments were selected in the following descending order of priority: (1) intracytoplasmic sperm injection, (2) in vitro fertilization, (3) intrauterine insemination, (4) ovulation induction, and (5) SC. 27 If this information was unavailable in the medical record transcripts, “blastocyst transfer” reported by mothers on the questionnaire was classified as in vitro fertilization. 27 Thus, while in vitro fertilization and intracytoplasmic sperm injection fell under the narrower definition of ART, ovulation induction and intrauterine insemination were categorized as non-ART, as used by the US Centers for Disease Control and Prevention and the American Society for Reproductive Medicine. 2 The following variables reported to be associated with fertility treatments or KD in previous studies were included as covariates in the present study 8 , 18 , 20 , 21 , 22 , 28 , 29 , 30 ; birth year (2011, 2012, 2013, or 2014, and missing); child’s sex (male or female, and missing); gestational length at birth (preterm, term, or postterm, and missing); multiple pregnancies (yes or no); maternal age at delivery or birth (<25, 25-29, 30-34, 35-39, or ≥40 years, and missing); maternal history of endometriosis (yes or no), polycystic ovary syndrome (yes or no), allergic rhinitis (yes or no), atopic dermatitis (yes or no), asthma (yes or no), and autoimmune diseases (yes or no); older sibling (yes or no); household income (<¥200 000; ¥200 000-¥399 000; ¥400 000-¥599 000; or ≥¥600 000 [to convert to US dollars, multiply by 0.0062]); and area of residence at registration (15 areas based on Regional Centres). The maternal history of these conditions was derived from a self-administered questionnaire completed at study registration. For autoimmune diseases, given the small number of patients with each individual condition recorded (collagen disease, autoimmune disease, systemic lupus erythematosus, rheumatoid arthritis, and other collagen disease or immune-related disease), a composite variable of autoimmune disease representing the presence of any autoimmune disease was created. Household income was derived from a self-administered questionnaire during the second and third trimesters. Other items were transcribed from medical records by physicians, midwives, or nurses, and/or research coordinators. Descriptive statistics were presented as counts and proportions for all variables. Person-years were calculated from birth to the first end point: the KD onset event or the last questionnaire response, whichever came first. When there was no questionnaire response regarding the child, we set the follow-up period to 1 month if maternal medical transcription data at 1 month post partum was available, or to 0 months if only live birth confirmation was available without any maternal data. Incidence rates were calculated by dividing the number of cases of KD by the total person-years of follow-up, expressed per 100 000 person-years. Hazard ratios (HRs) and corresponding 95% CIs were calculated using Cox proportional hazards regression models after adjusting for potential confounding factors. To account for clustering of siblings from the same mother, cluster-robust variance estimation by mother ID was applied to all models. In model 1, we adjusted for birth year, child’s sex, gestational length at birth, multiple pregnancies, maternal age, older sibling, household income, and area. In model 2, we further adjusted for maternal endometriosis and maternal polycystic ovary syndrome, and in model 3 we additionally adjusted for maternal history of allergic rhinitis, atopic dermatitis, asthma, and autoimmune disease. For exposure variables, we calculated HRs for AC using SC as the reference and for fertility treatment (ovulation induction, intrauterine insemination, in vitro fertilization, and intracytoplasmic sperm injection) with SC as the reference. The assumption in the Cox proportional hazards regression analysis was assessed by visual inspection of log (-log [survival]) plots and by examination of scaled Schoenfeld residuals, and it was not violated. The same analytical methods were applied as in the main analysis. Missing values were included as dummy variables. Tests of statistical significance were based on 2-tailed P values or 95% CIs, and statistical significance was set at P  < .05. SAS software, version 9.4 (SAS Institute Inc) was used for the statistical analyses.

Among 104 043 pregnancies, there were 102 407 live births. We excluded those with missing information on the conception method (n = 543), leaving a final sample size of 101 864 (eFigure in Supplement 1 ). Among the 101 864 participants (mean [SD] gestational age, 38.8 [1.7] weeks; 51 402 boys [50.5%] and 48 864 girls [48.0%]), 93 484 (91.8%) were born from SC and 8380 (8.2%) from AC, which were further divided into 3381 cases of ovulation induction (3.3% of all participants), 1536 cases of intrauterine insemination (1.5%), 1619 cases of in vitro fertilization (1.6%), 1798 cases of intracytoplasmic sperm injection (1.8%), and 46 cases of ART with missing details (0.1%) ( Table 1 ). AC (either non-ART or ART) was associated with preterm births, multiple births, and older mothers. Those who underwent non-ART, particularly ovulation induction, were more likely to have polycystic ovary syndrome, whereas those who underwent ART (in vitro fertilization and intracytoplasmic sperm injection) were more likely to have endometriosis and higher household income compared with those who underwent SC. Abbreviations: AC, assisted conception (fertility treatments); ART, artificial reproductive technique; SC, spontaneous conception. To convert to US dollars, multiply by 0.0062. During the 431 558 person-years of follow-up for the 101 864 participants, there were a total of 1226 cases of KD before 5 years (≤59 months) of age ( Table 2 ). The incidence rates of KD per 100 000 person-years were 284 overall: 274 for SC and 391 for AC. Compared with SC, AC was associated with an increased risk of KD after adjusting for covariates in model 1 (adjusted HR [AHR], 1.42 [95% CI, 1.18-1.70]). When AC was divided into non-ART and ART, the corresponding incidence rates per 100 000 person-years and adjusted HRs compared with SC were 429 (AHR, 1.54 [95% CI, 1.24-1.92]) for non-ART and 340 (AHR, 1.23 [95% CI, 0.92-1.63]) for ART. When further divided, the incidence rates per 100 000 person-years were 437 for ovulation induction, 410 for intrauterine insemination, 438 for in vitro fertilization, and 246 for intracytoplasmic sperm injection. Ovulation induction (AHR, 1.57 [95% CI, 1.22-2.03]), intrauterine insemination (AHR, 1.48 [95% CI, 1.01-2.16]), and in vitro fertilization (AHR, 1.53 [95% CI, 1.07-2.20]) were associated with an increased risk of KD compared with SC, whereas intracytoplasmic sperm injection was not associated with the risk of KD- (AHR, 0.91 [95% CI, 0.58-1.43]). In model 2, which was further adjusted for endometriosis and polycystic ovary syndrome, the increased risks did not change; the AHR for ovulation induction was 1.55 (95% CI, 1.19-2.01), for intrauterine insemination was 1.46 (95% CI, 1.00-2.15), and for in vitro fertilization was 1.52 (95% CI, 1.06-2.19). In model 3, which was further adjusted for maternal allergic conditions and autoimmune disease, similar patterns were observed; the AHR for ovulation induction was 1.56 (95% CI, 1.19-2.03), for intrauterine insemination was 1.41 (95% CI, 0.95-2.09), and for in vitro fertilization was 1.55 (95% CI, 1.08-2.22). Abbreviation: ART, artificial reproductive technique. Adjusted for birth year, mother’s age, multiple pregnancies, gestational age at birth, child’s sex, older sibling, household income, and area. Adjusted further for mother’s endometriosis and polycystic ovary syndrome. Adjusted further for mother’s allergic rhinitis, atopic dermatitis, asthma, and autoimmune disease.

In the present nationwide prospective birth cohort study, fertility treatments were associated with an increased risk of KD until 59 months of age after adjusting for potential confounding factors such as birth year, child sex, and perinatal and maternal factors. Increased risks were similarly observed among ovulation induction, intrauterine insemination, and in vitro fertilization, but not intracytoplasmic sperm injection, and did not change materially after additional adjustment for maternal endometriosis and polycystic ovary syndrome. The observed incidence rate of KD in this study of children born in 2011 to 2014 (284 per 100 000 person-years) was slightly lower than the corresponding incidence reported by a hospital-based study from Japan (1444 pediatric departments in hospitals with ≥100 beds) conducted in 2015 and 2016 (309 per 100 000 population per year), 31 although the data from the 2 studies are not strictly comparable because of the difference in sampling. To our knowledge, this is the first study to assess whether fertility treatments are associated with KD among offspring. We found that, with the exception of intracytoplasmic sperm injection, fertility treatments were associated with increased risks of KD among children. The potential biological mechanisms for our findings are as follows. First, the ovarian stimulation used in most fertility treatments creates a supraphysiological hormonal environment during oocyte maturation and early embryonic development. 32 , 33 This hormonal environment has been reported to perturb 1-carbon cycle metabolism, leading to homocysteine accumulation in follicular fluid, a potent inhibitor of DNA methyltransferase activity. 34 Such disruption of methylation during early embryonic development may induce epigenetic predisposition in offspring. Second, growing evidence suggests that such ART procedures induce epigenetic modifications, including altered methylation patterns at imprinted genes in offspring, which may affect immune system development and its function. 35 , 36 , 37 In KD, hypomethylation due to immune regulatory genes including toll-like receptor family ( TLR1 , TLR2 , TLR4 , TLR6 , TLR8 , and TLR9 ) and FCGR2A was found in the acute phase, with upregulated expression of these genes leading the systemic inflammation. 38 , 39 KD is believed to result from an aberrant immune response to infectious triggers in genetically susceptible individuals, 14 and altered immune programming due to epigenetic modifications induced by ART could increase disease susceptibility. Increased risks of KD were observed among children conceived through conventional in vitro fertilization, intrauterine insemination, and ovulation induction, but not intracytoplasmic sperm injection. This differential association may be explained by the underlying cause of infertility. Conventional in vitro fertilization, intrauterine insemination, and ovulation induction are more commonly used for female factor infertility, including endometriosis and polycystic ovary syndrome, 40 , 41 whereas intracytoplasmic sperm injection is primarily indicated for male factor infertility. 42 Another consideration is that intracytoplasmic sperm injection involves artificial sperm selection and mechanical puncture of the oocyte membrane, whereas sperm penetrate the oocyte naturally without direct manipulation of the intracellular environment in conventional in vitro fertilization. 42 , 43 This procedural difference could theoretically result in different consequences, although the specific mechanisms remain unclear. Our finding that the increased risk of KD was significant even after adjusting for maternal endometriosis and polycystic ovary syndrome suggests that the association is not entirely explained by the maternal conditions that led to infertility, but rather that the fertility treatment procedures themselves or other unmeasured aspects of subfertility may play a role. This study has several strengths, including its nationwide prospective birth cohort design and long follow-up periods. The large sample size provided adequate statistical power to detect small effect sizes and examine different types of fertility treatments separately. Data collection through both medical record transcripts and parental self-reports enhanced the accuracy of our fertility treatment classifications. The use of a comprehensive national database minimized selection bias and ensured generalizability to the broader population. Furthermore, the ability to adjust for important maternal factors, including underlying infertility diagnoses and socioeconomic factors, strengthened the internal validity of our findings. However, several limitations should be acknowledged. First, despite the large overall sample size, statistical power may have been insufficient to detect associations in smaller subgroups, such as intracytoplasmic sperm injection–conceived children. Second, the possibility of some misclassification of fertility treatment methods cannot be excluded, despite the use of medical record data. In addition, information on specific treatment protocols, including details of ovarian stimulation regimens, culture media compositions, and embryo quality, was not available in our dataset. Third, some cases may have been undetected if they were not reported by the parents in the questionnaires. Fourth, we did not account for unmeasured variables, including genetic factors.

In this nationwide cohort study, most fertility treatments, with the exception of intracytoplasmic sperm injection, were associated with an increased risk of KD among offspring. Further research is necessary to examine the consistency of our findings and elucidate the underlying mechanisms.

Infertility affects approximately 17.5% of couples of reproductive age worldwide. 1 Various fertility treatments have been developed, ranging from ovulation induction with medications to more complex laboratory-based procedures such as in vitro fertilization and intracytoplasmic sperm injection. 2 , 3 , 4 The use of fertility treatments worldwide has increased markedly over the past several decades. 5 , 6 In 2023, approximately 11.7% of all live births in Japan resulted from assisted reproductive technology (ART), 4 , 7 which is substantially higher than the rates in Europe (7.9%) and the US (5.1%). 8 Given this widespread and increasing use of ART, there is also growing interest in understanding the long-term health outcomes of children conceived by these technologies. Previous reports suggest that ART procedures may induce epigenetic modification, including alterations in DNA methylation patterns, 9 which has been observed in Kawasaki disease (KD). 39 According to a large birth cohort study in Norway, ART–conceived offspring showed DNA methylation in 176 genes, including genes related to growth, neurodevelopment, and other health outcomes. 10 KD is a type of systemic vasculitis that primarily affects young children. It is characterized by a severe inflammatory response 11 , 12 and was first reported in 1967 by Kawasaki. 13 Although the exact cause remains unclear, infectious agents are thought to play a role in triggering KD, 12 , 14 which remains a leading cause of acquired heart disease in children. The annual incidence rate of KD per 100 000 children aged 59 months or younger in Japan has increased markedly, from 100 in 2001 to 371 in 2019. 15 , 16 Approximately 20% of patients not properly treated develop coronary artery aneurysms after KD, leading to lifelong cardiac sequelae. 14 , 15 , 17 Known risk factors for the onset of KD include male sex, age younger than 12 months, presence of siblings, preterm birth, higher socioeconomic status factors, mother’s immune disorders, genetic susceptibility, and certain environmental exposures such as ambient ozone levels and extreme heat. 14 , 18 , 19 , 20 , 21 , 22 , 23 Despite extensive research on ART and the health of offspring, to our knowledge, no studies have examined whether fertility treatment is associated with KD, both of which are considered to involve immune-related mechanisms and have shown increasing trends. 1 , 4 , 14 , 22 Therefore, we conducted a nationwide prospective birth cohort study to examine whether conception through various fertility treatments, including ovulation induction, intrauterine insemination, in vitro fertilization, and intracytoplasmic sperm injection, is associated with an increased risk of developing KD during early childhood. We hypothesized that children born after ART would have an increased risk of developing KD.