Association between maternal infertility treatment and child neurodevelopment: findings from the Tohoku Medical Megabank Project Birth and Three-Generation Cohort Study in Miyagi and Iwate Prefectures, Japan

Abstract

Objectives

This study aimed to examine the association between infertility treatment and neurodevelopment in children at 2 and 3.5 years of age.

Design

Prospective cohort study.

Setting and participants

The study population consisted of mother–child pairs who participated in the Tohoku Medical Megabank Project Birth and Three-Generation Cohort Study in Miyagi and Iwate Prefectures, Japan. Pregnant women were recruited in obstetric clinics or hospitals and their children were followed up by the questionnaire.

Outcome measures

The children’s neurodevelopmental outcomes were assessed at 2 and 3.5 years of age using the Ages and Stages Questionnaire, third edition (ASQ-3), which consists of questions on five developmental domains. We performed a multivariate logistic regression analysis of the association between infertility treatment (including ovulation induction (OI), artificial insemination with husband’s sperm (AIH) and assisted reproductive technology (ART)) and the clinical range of ASQ-3.

Results

Of 9655 mother–child pairs, 273 (2.8%) and 487 (5.0%) were conceived through OI/AIH and ART, respectively. The odds of having developmental delays at 2 years of age were higher in children conceived through OI/AIH (OR, 1.36; 95% CI 1.00 to 1.85) and ART (OR, 1.36; 95% CI 1.07 to 1.72) than in those conceived naturally. Additionally, OI/AIH and ART were significantly associated with communication (OR, 1.93; 95% CI 1.25 to 2.98) and gross motor (OR, 1.50; 95% CI 1.08 to 2.09) delays, respectively. There were no statistically significant differences in the odds of having developmental delays at 3.5 years of age in children conceived through OI/AIH (OR, 1.13; 95% CI 0.79 to 1.61) and ART (OR, 1.03; 95% CI 0.78 to 1.37).

Conclusion

In this study, we found a significant association between infertility treatment and children’s neurodevelopment at 2 years of age, whereas no statistically significant differences were found at 3.5 years of age.

Strengths and limitations of this study.

• The large sample size in this study allowed us to investigate the association between each type of infertility treatment and neurodevelopmental outcomes.

• We evaluated children’s neurodevelopment at two points—2 and 3.5 years of age.

• Ages and Stages Questionnaire, third edition is a screening tool (not a diagnostic tool) for developmental delays.

• The number of children born through specific forms of infertility treatment, such as fresh embryo transfer, was insufficient to allow subanalysis of fresh or frozen embryo transfer.

Introduction

With the rapid progress in infertility treatment, the number of women who use infertility treatment is increasing, and more than 10% of the childbearing population has resorted to assisted reproductive technology (ART) for conception.¹ In recent years, the number of fertility treatments in Japan has also been on the increase. In 2018, 56 000 newborns were conceived through ART, including in vitro fertilisation (IVF) and intracytoplasmic sperm injection (ICSI). This accounts for approximately 6% of all live births in the same year.² Because conception using ART involves several processes (such as physical manipulation of gametes and exposure of embryos to hormones and culture media) that differ greatly from those in a natural conception, there is potential for disturbance of normal early developmental processes.^{3 4} Therefore, many studies have investigated the association between ART and neurodevelopmental outcomes.^5–8

The findings have been almost entirely consistent in showing that singleton children conceived through ART and born at term are no different neurodevelopmentally from those born following natural conception. However, the evidence remains equivocal,^8–11 and another review stated that possible associations between infertility treatment and developmental delay require further assessment in larger studies.¹² The inconsistent results in the previous studies may have been due to the small sample size (fewer than 1000 children),³ and a longitudinal study with a larger sample size is needed.

This study aimed to investigate the association between infertility treatment and children’s neurodevelopment at 2 and 3.5 years of age in a Japanese birth cohort.

Methods

Study setting and participants

This study was based on data obtained by the Tohoku Medical Megabank Project Birth and Three-Generation Cohort Study (TMM BirThree Cohort Study). The TMM BirThree Cohort Study is a prospective cohort study based in Miyagi and Iwate Prefectures, Japan and has been published elsewhere.^{13 14} Pregnant women and their family members were contacted in obstetric clinics or hospitals from 2013 to 2017, and 23 406 pregnant women participated in the study. Written informed consent was obtained from all participants. All participants were free to decline consent to participate in the research and were told that there was no disadvantage or risk involved with their refusal to participate. The TMM BirThree Cohort Study protocol was approved by the Tohoku University Tohoku Medical Megabank Organization’s internal review board (2013-1-103-1).

Infertility treatments

We extracted infertility treatment types from maternal medical records, and they included ovulation induction (OI), artificial insemination with husband’s sperm (AIH), IVF and ICSI; we refer to IVF and ICSI as ART in this study.

Neurodevelopmental assessments

The third edition of the Ages and Stages Questionnaire (ASQ-3) was used to evaluate child neurodevelopment.^{15 16} ASQ-3 is a comprehensive, reliable screening questionnaire that can be used for children aged 1 and 66 months. The guardians of children aged 2 and 3.5 years filled in the questionnaire. Each questionnaire contains 30 questions divided into five developmental domains: ‘communication’, ‘fine motor’, ‘gross motor’, ‘problemsolving’ and ‘personal-social’. Each domain has a set of six items, and each item is given a score of 10, 5 and 0 corresponding to ‘yes’, ‘sometimes’ and ‘not yet’, respectively. The total score ranges from 0 to 60 for each domain. We defined ‘developmental delay’ in a domain as when the score was greater than two SD below the mean in that domain.¹⁷ In this study, we used the validated Japanese translation of ASQ-3.¹⁶

Covariates

Considering previous studies and the characteristics of the population in this study, we included maternal age at delivery, parity, preterm birth (PTB), child sex, birth defects, multiple births, maternal education and household income as covariates of developmental outcomes.^{7 18} Maternal age, gestational weeks, parity, child sex, birth defect and multiple birth were obtained from medical records. Birth defects were defined in this study as follow: anencephaly, microcephaly, hydrocephalus, craniotabes, holoprosencephaly, agenesis of the corpus callosum, other head or brain abnormalities, omphalocele, abdominal fissure, epidermolysis bullosa hereditarian, incontinentia pigmenti, myelomeningocele, Down’s syndrome, trisomy 18, trisomy 13, achondroplasia, osteogenesis imperfecta, arthrogryposis multiplex congenita, other skeletal or muscle abnormalities, amniotic band syndrome and other chromosomal abnormality. Data on maternal level of education and household income were obtained using a self-report questionnaire.

Statistical analysis

Characteristics of mothers and children were compared in three groups: natural conception, OI/AIH and ART. Continuous and categorical variables were described as mean (SD) and frequency or proportion, respectively. Differences in prevalence were analysed using the χ² test. Logistic regression analyses were performed to determine the associations between infertility treatment and each of the five domains of ASQ-3 at 2 and 3.5 years of age after adjusting for possible confounding factors. ORs and CIs were calculated. Children who were conceived naturally were used as the reference group in all analyses. For subgroup analyses, participants were classified into five groups as follows: natural conception, OI, AIH, IVF and ICSI. Furthermore, we compared developmental outcomes among children born following natural conception, fresh embryo transfer (ET) and frozen-thawed ET. Statistical significance was set at p<0.05. All statistical analyses were performed using SAS V.9.4 (SAS Institute, Cary, North Carolina).

Patient and public involvement

No participants were involved in this study because we used an existing data set for the analysis.

Results

Characteristics of the participants

We analysed data from 9655 eligible mother–child pairs who filled in ASQ-3 questionnaires (figure 1). Of the 9655 mother–child pairs, 273 (2.8%) were conceived through OI/AIH and 487 (5.0%) through ART. Maternal and child characteristics are shown in table 1. Mothers who were exposed to infertility treatment were older and had higher levels of educational and household income. Children who were conceived following infertility treatment were likely to have a lower birth weight (LBW) and gestational age.

Figure 1.

Flowchart of the exclusion criteria in the TMM BirThree Cohort Study. This flowchart describes the exclusion criteria and the number of total participants, excluded participants, and eligible participants. ASQ-3, Ages & Stages Questionnaires, Third Edition; TMM BirThree Cohort Study, Tohoku Medical Megabank Project Birth and Three-Generation Cohort Study.

Table 1.

Characteristics of the study population

	Total		Infertility treatment						P
			Natural conception		OI/AIH		ART
	n=9655	%	n=8895	%	n=273	%	n=487	%
Mother’s age, mean (SD), years	32.7	4.7	32.4	4.6	34.3	3.9	37.2	3.8	<0.0001
Pre-pregnancy BMI (kg/m2)
<18.5	1365	14.1	1279	14.4	36	13.2	50	10.3	0.0707
18.5 to <25.0	7204	74.6	6636	74.6	196	72.1	372	76.3
≧ 25.0	1083	11.2	977	11.0	41	15.0	65	13.4
Educational level
High school graduate or less	2711	28.1	2565	28.8	49	18.0	97	19.9	<0.0001
Junior or vocational college graduate	3513	36.4	3215	36.1	100	36.6	198	40.7
University graduate or above	2722	28.2	2457	27.6	102	37.4	163	33.5
Household income (JPY/year)
<4 000 000	3117	32.3	2974	33.4	63	23.1	80	16.4	<0.0001
4 000 000 to <6 000 000	3063	31.7	2810	31.6	86	31.5	167	34.3
≥6 000 000	2992	31.0	2658	29.9	111	40.7	223	45.8
Cigarette smoking
Never	6194	64.6	5675	64.3	195	71.4	324	66.7	<0.0001
Stopped before pregnancy	2291	23.9	2075	23.5	60	22.0	156	32.1
Stopped after pregnancy	950	9.9	928	10.5	17	6.2	5	1.0
Smoked during early pregnancy	152	1.6	150	1.7	1	0.4	1	0.2
Alcohol consumption
Drinking at early pregnancy	1917	20.0	1766	20.0	57	21.0	94	19.3	0.5426
Former	3273	34.1	3023	34.2	93	34.2	157	32.3
Never	3871	40.3	3568	40.4	105	38.6	198	40.7
Cannot drink because of constitution	536	5.6	482	5.5	17	6.3	37	7.6
Parity
Nullipara	4564	47.3	4024	45.2	202	74.0	338	69.4	<0.0001
Multipara	5091	52.7	4871	54.8	71	26.0	149	30.6
Child’s sex
Male	4958	51.4	4581	51.5	137	50.2	240	49.3	0.5874
Female	4697	48.7	4314	48.5	136	49.8	247	50.7
Gestational age (weeks)
≥37	9017	93.4	8338	93.7	245	89.7	434	89.1	<0.0001
34 to <37	482	5.0	417	4.7	26	9.5	39	8.0
29 to <34	113	1.2	98	1.1	2	0.7	13	2.7
<29	33	0.3	32	0.4	0	0.0	1	0.2
Birth weight (g)
≥2500	8689	90.0	8037	90.4	229	83.9	423	86.9	0.0004
<2500	952	9.9	844	9.5	44	16.1	64	13.1
Birth defect
No	9477	98.3	8740	98.4	269	98.5	468	96.1	0.0033
Yes	169	1.8	146	1.6	4	1.5	19	3.9
Multiple birth
No	9440	97.8	8746	98.3	244	89.4	450	92.4	<0.0001
Yes	215	2.2	149	1.7	29	10.6	37	7.6
ASQ-3 (2Y)
>Mean −2SD	8218	85.1	7626	85.7	215	78.8	377	77.4	<0.0001
≤Mean −2SD	1437	14.9	1269	14.3	58	21.3	110	22.6
ASQ-3 (3.5Y)
>Mean −2SD	8398	87.0	7746	87.1	233	85.4	419	86.0	0.5749
≤Mean −2SD	1257	13.0	1149	12.9	40	14.7	68	14.0
Father’s age, mean (SD), y	n=4183		n=3784		n=142		n=257
	34.5	5.7	34.2	5.6	35.1	4.8	38.8	5.2	<0.0001

AIH, artificial insemination with husband’s semen; ART, assisted reproductive technology; ASQ-3, Ages & Stages Questionnaires, Third Edition; BMI, body mass index; OI, ovulation induction.

Infertility treatment and children’s neurodevelopment at 2 years of age

In total, 1437 children (14.9%) had developmental delay at 2 years of age when screened using the ASQ-3. The proportion of children with developmental delays at 2 years of age was 14.3%, 21.3% and 22.6% for natural conception, OI/AIH and ART, respectively. In the multivariable models, the odds of having developmental delays at 2 years of age were higher in children conceived through OI/AIH (OR, 1.36; 95% CI 1.00 to 1.85) and ART (OR, 1.36; 95% CI 1.07 to 1.72) than in those conceived naturally. OI/AIH and ART were significantly associated with communication (OR, 1.93; 95% CI 1.25 to 2.98) and gross motor (OR, 1.50; 95% CI 1.08 to 2.09) delays, respectively (table 2).

Table 2.

Odds ratios (95% CIs) from logistic regression models for infertility treatment and children’s neurodevelopment at the age of 2 and 3.5 years (n=9655)

	N	2 years								3.5 years
		n	%	Crude			Adjusted			n	%	Crude			Adjusted
				OR	95% CI		OR	95% CI				OR	95% CI		OR	95% CI
					Lower	Upper		Lower	Upper				Lower	Upper		Lower	Upper
Total score
Natural conception	8895	1269	14.27	ref			ref			1149	12.92	ref			ref
OI/AIH	273	58	21.25	1.62	1.21	2.18	1.36	1.00	1.85	40	14.65	1.16	0.82	1.63	1.13	0.79	1.61
ART	487	110	22.59	1.75	1.41	2.19	1.36	1.07	1.72	68	13.96	1.09	0.84	1.42	1.03	0.78	1.37
Communication
Natural conception	8895	445	5.00	ref			ref			482	5.42	ref			ref
OI/AIH	273	27	9.89	2.09	1.39	3.14	1.93	1.25	2.98	17	6.23	1.16	0.70	1.91	1.13	0.68	1.90
ART	487	27	5.54	1.12	0.75	1.66	0.95	0.62	1.45	29	5.95	1.11	0.75	1.63	1.04	0.69	1.56
Gross motor
Natural conception	8895	483	5.43	ref			ref			367	4.13	ref			ref
OI/AIH	273	20	7.33	1.38	0.87	2.19	1.15	0.72	1.85	16	5.86	1.45	0.86	2.42	1.26	0.74	2.13
ART	487	49	10.06	1.95	1.43	2.65	1.50	1.08	2.09	25	5.13	1.26	0.83	1.91	1.03	0.67	1.60
Fine motor
Natural conception	8895	389	4.37	ref			ref			513	5.77	ref			ref
OI/AIH	273	17	6.23	1.45	0.88	2.40	1.28	0.76	2.13	19	6.96	1.22	0.76	1.97	1.20	0.74	1.96
ART	487	34	6.98	1.64	1.14	2.36	1.27	0.86	1.86	31	6.37	1.11	0.76	1.62	1.05	0.71	1.56
Problem solving
Natural conception	8895	354	3.98	ref			ref			505	5.68	ref			ref
OI/AIH	273	13	4.76	1.21	0.68	2.13	0.90	0.50	1.60	18	6.59	1.17	0.72	1.91	1.20	0.73	1.99
ART	487	32	6.57	1.70	1.17	2.47	1.18	0.79	1.75	27	5.54	0.98	0.66	1.45	0.91	0.60	1.38
Personal-social
Natural conception	8895	449	5.05	ref			ref			400	4.5	ref			ref
OI/AIH	273	24	8.79	1.81	1.18	2.79	1.46	0.93	2.28	16	5.86	1.32	0.79	2.21	1.20	0.70	2.04
ART	487	42	8.62	1.78	1.28	2.47	1.25	0.88	1.77	28	5.75	1.30	0.87	1.92	1.05	0.69	1.60

Multivariable logistic models were adjusted for maternal age, parity, gestational week (<37 weeks), child sex, birth defect, multiple birth, maternal education level and household income.

AIH, artificial insemination with husband’s semen; ART, assisted reproductive technology; OI, ovulation induction.

Infertility treatment and children’s neurodevelopment at 3.5 years of age

In total, 1257 children (13.0%) had developmental delay at 3.5 years of age when screened using the ASQ-3. The proportion of children with developmental delay at 3.5 years of age was 12.9%,14.7% and 14.0% for natural conception, OI/AIH and ART, respectively. In the multivariable models, the odds of having developmental delays at 3.5 years of age were higher in children conceived through OI/AIH (OR, 1.13; 95% CI 0.79 to 1.61) and ART (OR, 1.03; 95% CI 0.78 to 1.37) than in those conceived naturally. Moreover, children conceived through infertility treatment had no statistically significant differences in any domains at 3.5 years of age (table 2).

Subgroups based on the five domains and embryo types were analysed and the results are shown in online supplemental tables S1-S4. Although the number of children conceived through each infertility treatment was less, the results, when compared within the five groups or types of ET, were similar to those within the three groups.

Discussion

We investigated the association between infertility treatment and children’s neurodevelopment, among 9655 mother–child pairs. In multivariable models, a significant association between infertility treatment and neurodevelopment was observed among the children conceived through infertility treatment, at 2 years of age; no statistically significant differences were found in them at 3.5 years of age.

Children conceived through ART are known to be at risk of LBW, PTB and birth defects,¹⁹ which are risk factors for neurodevelopmental disorders.^{20 21} However, the data from previous studies among children aged 1–5 years suggest that there are no developmental differences between children conceived through ART and those conceived naturally, after adjusting for confounding variables.^{3 19 22} Even when children with congenital anomalies or genetic syndromes and multiple births (which are known to affect development) were excluded from the study (n=9271), a significant association between infertility treatment and children’s neurodevelopment was seen at 2 years of age, while no statistically significant differences were found at 3.5 years of age (online supplemental table S5). Furthermore, we took into account very preterm (<34 weeks) or extremely preterm (<29 weeks) birth for adjusting the results for important risk factors and the similar result was obtained (online supplemental table S6). We also performed analysis adjusted for birth weight and the similar result was obtained (online supplemental table S7). A systematic review of neurodevelopmental disorders,⁵ neuromotor, cognitive, language, and behavioural outcomes of children born after ART showed that ART had no significant effect on children’s neuromotor and cognitive development.⁶ A previous study showed no significant difference in the neurodevelopmental scores between children conceived through ART and those conceived naturally; however, it showed a decrease in the score units of each scale among children conceived through ART.²³ The reason for the discrepancy in results at 2 years of age between our study and the previous study might be the proportion of children conceived through infertility treatment. Because infertility treatment has become widespread owing to technological advancement, the proportion of children conceived through infertility treatment was higher in our study than in previous studies.^24–26 Therefore, statistical differences might be detectable in our study. In addition, frozen-thawed embryos have been used for most ARTs in Japan.²⁷ Frozen-thawed ET is associated with a significantly lower incidence of PTB and LBW, which are risk factors for neurodevelopmental disorders.²⁸ Among children with developmental delay at 2 years of age, the proportion with developmental delay at 3.5 years of age was 47.3%,46.6% and 38.5% for natural conception, OI/AIH and ART, respectively. The proportion of children with developmental delay at 3.5 years of age was low, especially in those conceived through ART. Our longitudinal study demonstrated that at 2 and 3.5 years of age, the neurodevelopmental status of children conceived through infertility treatment, especially ART, might eventually catch up with that of those conceived naturally. In Japan, the health check-up for children is done at 1.5 years of age, so the counsel provided by doctors and public health nurses, which is tailored towards individual development, might also help promote the children’s development. A previous study in Japan stated that the developmental state of a child needs to be understood not only based on the characteristics of the child at one point but also on the progress of the child’s development from the results of the health check-up at 1.5 years of age. In fact, it has been suggested that it is important to conduct a follow-up health check-up at 3 years of age in relation to the development at the previous age.²⁹

Developmental delay in the communication domain was seen at age 2 years in children conceived through OI/AIH but not in those conceived through ART. A previous study showed that compared with children conceived naturally, those conceived through OI had lower verbal ability at 3 years of age, while those conceived through ART had higher verbal ability; children conceived through infertility treatment had higher verbal ability at 5 years of age than those conceived naturally.³⁰ A previous hospital-based cohort study in Japan showed that children conceived through ART had significantly better language development than those conceived naturally.⁷ Although direct comparisons with other studies assessing neurodevelopment in children conceived through infertility treatment are difficult because each researcher uses various assessment tools and control groups, the children conceived through ART might have higher verbal ability or better language development than those conceived through OI/AIH or naturally. Children conceived through infertility treatments generally enjoy advantageous socioeconomic benefits. It is possible that parents who use ART to conceive may invest more into rearing the children, which might have the greatest impact on language skills.^{7 30 31}

Furthermore, ART was associated with developmental delays in the gross motor domain at 2 years of age. This association was not seen at 3.5 years of age. A hospital-based cohort study in Japan reported no significant difference in gross motor function (evaluated using the Ability for Basic Movement Scale for Children) at 12 months of age between the ART and control groups.³² In contrast, the Danish National Birth Cohort study showed that infertility treatment, especially ICSI, may be associated with a slight delay in gross motor development at 1.5 years of age.⁸ However, another study suggested no differences in motor development at 5 years of age among children conceived through ICSI, IVF and natural conception.³³ These findings suggest that infertility treatment might be associated with children’s development in the early stage, but the neurodevelopment of children conceived through infertility treatment might catch up with that of those conceived naturally.

This study had some limitations. The ASQ-3 is a screening tool (not a diagnostic tool) for developmental delays; however, it is considered to be highly reliable because it has been validated in many countries worldwide and has been used in a variety of studies.^{18 34 35} Since we focused on children who filled in both questionnaires (at 2 and 3.5 years of age), there was the concern of bias due to the reduced number of participants and the deterioration of statistical power in the analysis. However, among the children who did not fill in the questionnaire at 3.5 years of age, the percentage of those who had neurodevelopmental delay at 2 years of age was 14.5%,10.8% and 15.0% for natural conception, OI/AIH and ART, respectively. Therefore, there was no bias such as more children with neurodevelopmental delay at 2 years of age among children who did not answer the questionnaire at 3.5 years old and the potential applicability of our results to national populations can be discussed. In the subgroup analysis with detailed classification, the number of children conceived through specific forms of infertility treatment, such as fresh ET, was insufficient to compare the ET types. Male infertility or the time to pregnancy should be taken into account when evaluating subfertility patients.³⁶ However, it was difficult to include these factors because we did not collect them in this study. Research in this field is complex because of the need to collect various data to determine the effects of infertility treatment on offspring outcomes. As this was an observational study, residual confounding might have occurred. However, this study was a large longitudinal birth cohort with detailed information from the participants, and although continued follow-up of children born after infertility treatment is needed, this study helps to increase the understanding of the association between infertility treatment and neurodevelopmental outcomes in Japanese children.

Conclusion

In this study, we found a significant association between infertility treatment and neurodevelopment among children who were conceived through infertility treatments and those conceived naturally, at 2 years of age; however, no statistically significant differences were found at 3.5 years of age.

Data availability statement

Data are available upon reasonable request. The data obtained through the TMM BirThree Cohort Study are incorporated into the TMM Biobank. All data analysed during the present study are available for research purpose with the approval by the Sample and Data Access Committee of the Biobank.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

The TMM BirThree Cohort Study protocol was approved by the Tohoku University Tohoku Medical Megabank Organization’s internal review board (2013-1-103-1). Participants gave informed consent to participate in the study before taking part.