The potential health risks of exposure to environmental chemicals ― Global implications for future generations

Abstract

In 2001, we launched the Hokkaido Study, the first prospective birth cohort study in Japan. We are currently tracking the effects of environmental chemicals, using a life course approach. The study examines life circumstances after birth, and the longest follow-up to date is 20 years of age. We have measured prenatal exposure to dioxins, organochlorine pesticides, per- and polyfluoroalkyl substances, plasticizers such as di(2-ethylhexyl) phthalate, and bisphenol A. Our findings have mostly revealed that increased exposure to these environmental chemicals is linked to increased risk of lower birth size, effects on thyroid and steroid hormones, adipokine levels, as well as disruption of neurodevelopment, including causing asthma and respiratory symptoms. However, it should be noted that our findings also include protective or null findings, which may be due to low chemical concentrations or differences in prenatal or postnatal exposure. We would like to emphasize the importance of long-term continuation of the cohort, effective utilization of the data, and application of the results to environmental and health policies.

1. Introduction

In recent years, environmental pollution on a global scale has greatly affected the survival of many living organisms, including humans. Since the Industrial Revolution, many chemical substances have been produced and their benefits are indispensable for human beings. However, there are already more than 275 million chemical substance registrations, and thousands more continue to be registered and updated every day.¹⁾ Data-based estimates from a European Union-supported publication suggest that global pollution kills 9 million people worldwide each year as a result of health hazards due to exposure to air, water, soil, and the workplace.²⁾ In addition, waste disposal of electrical and plastic products is “exported” to developing countries.³⁾ On the other hand, the Planetary Boundaries indicator for chemical pollution includes persistent organic pollutants (POPs), plastics, and heavy metals. POPs, in particular, once used, spread globally and have been detected in environmental and biological samples in uninhabited Arctic ice sheets, deserts, and deep oceans.⁴⁾ Mercury accumulation in wild animals has been increasing since the Industrial Revolution. In addition, there is no safe threshold for developmental neurotoxicity, especially for the heavy metal lead, and the World Health Organization (WHO) is particularly concerned about lead exposure in children.⁵⁾

This widespread chemical pollution, which not only threatens human health but also may cause a loss of biological diversity, is mainly attributed to the policies of industrialized countries, which prioritize economic growth over environmental conservation.⁶⁾ One direct cause is the quantitative increase in the use of various chemicals. For example, the dramatic increase in the use of chemical fertilizers and pesticides (insecticides, herbicides, fungicides, etc.) to increase agricultural productivity has drastically reduced the number of microorganisms, earthworms, bees, and butterflies that live in soil and water bodies, resulting in ecosystem monotonization.^7),8) In the years from 1970 to 2016, populations of fish, mammals, reptiles, birds, insects, etc., have been reported to have declined by about 70％.⁹⁾ It has already been pointed out that, among the various burdens placed on the global environmental system, chemical pollution is caused by the economic system of mass production, mass consumption, and mass disposal. In other words, the original pollution problem is now widely recognized as “global environmental pollution” and a “potential health risk problem.” Global warming and biodiversity loss have been treated as “two sides of the same coin,” and global environmental pollution as their “root” issue.

In 1996, “Our Stolen Future” was published by Theo Colborn et al.¹⁰⁾ In particular, the endocrine-disrupting effects of environmental chemicals have attracted great interest, and research on them has begun in many countries.¹¹⁾ Initially, the issues addressed were polychlorinated biphenyls (PCBs) and contaminated areas. Importantly, even low doses and minimal exposure levels of endocrine-disrupting chemicals can have significant effects, especially during critical development periods.¹¹⁾ These may not only appear in the exposed individuals but may also affect the next generation, highlighting the need for prospective cohort studies to fully understand the long-term effects of exposure to pollutants.

A cohort study, as depicted in Fig. 1, is a type of observational study that examines the association between factors and disease occurrence by following exposed and unexposed populations for a period of time and comparing the incidences of the disease under study. Studies based on a certain base (region, occupation, etc.) can be divided into two main types: cross-sectional studies, in which a single survey is conducted, and longitudinal studies, in which two or more surveys of the same subjects are conducted. In the latter, in particular, the first group of subjects surveyed is called a cohort. Cohort studies follow a population prospectively, allowing the process from exposure to disease development to be observed over time. The temporal sequence and logical flow of observations is similar to that of an experiment, and it is possible to study multiple diseases. It has the disadvantage of requiring large cohorts to be followed for long periods of time and cost when the occurrence of the disease under study is rare.

Fig. 1.

Positioning of cohort studies in epidemiological research.

2. Objectives and characteristics of the Hokkaido Study on environment and children’s health

In Japan, there were actually no birth cohort studies prospectively examining the effects of environmental exposures during the fetal period until 2000. We began epidemiological studies on genetic and environmental factors related to hypospadias and cryptorchidism in the 1990s, but due to the limitations of case-control studies in assessing exposure during pregnancy,¹²⁾ we launched a prospective study, “The Hokkaido Study on Environment and Children’s Health: Malformation, Development and Allergy (hereafter the Hokkaido Study)” in 2001.^13)-16)

“The Hokkaido Study” consists of two prospective cohorts, one large and one small. It is characterized by (1) assessment of the effects of exposure to low concentrations of environmental chemicals in daily life, (2) accurate measurement and analysis of environmental exposure factors during fetal life, including the period of fetal organogenesis, by collecting and storing maternal blood and umbilical cord blood during pregnancy, (3) longitudinal risk assessment of various postnatal outcomes such as congenital abnormalities, body size at birth, growth and development, allergic infections, thyroid and reproductive hormone effects, developmental disorders, and puberty development, (4) quantitative assessment of environmental factors (e.g., folic acid and cotinine levels in maternal blood during pregnancy), (5) identification of genetic high-risk groups with consideration of chemical metabolism and susceptibility to disease in the child, and (6) investigation of the mechanisms of next-generation effects, including comprehensive epigenomic effects.

Study procedures for the two cohorts are shown in Table 1. The Sapporo cohort consists of 514 mother-infant pairs of pregnant women who visited a single gynecological hospital in Sapporo. For this cohort, we are conducting detailed face-to-face neurobehavioral developmental examinations, and studying infant anthropometric growth, allergies and infectious diseases. Environmental chemical concentrations were measured during late pregnancy (23-35 of weeks gestation). The larger cohort, the Hokkaido cohort, consists of 20,926 mother-child pairs recruited in 41 community-based hospitals and clinics. In the Hokkaido cohort, environmental chemical concentrations were measured in pregnant women who gave consent by the 12th week of early pregnancy (organogenesis). The direct causal effects of exposure on birth defects, preterm birth, small for gestational age (SGA) and low birth weight, immunological and hormonal effects, and neurodevelopment (such as autism spectrum disorder and attention deficit hyperactivity disorder [ADHD]) of the offspring were evaluated. Currently, the oldest child in the Hokkaido cohort is 20 years old, while the youngest is 11 years old (as of December 2024).

Table 1.

Study procedures of the Sapporo cohort and Hokkaido cohort

	Questionnaire surveys			Face-to face-exams	Specimen/sample collection
	Neurobehavioral development	Allergies/infections	Anthropometric measurements/puberty
Sapporo cohort
6-7 months	EES			BSID-II, FTII,
1.5 years	EES	ISAAC, ATS-DLD, infections	Physical growth	BSID-II, DDST,
3.5 years	CBCL, EES	ISAAC, infections	Physical growth	K-ABC, WAIS-R,
7 years	CBCL, J-PSAI, 2D/4D	ISAAC, infections	Physical growth	WISC-III, WCST-KFS,
12 years	Tanner staging, onset		Tanner staging, onset of puberty	Event-related brain potentials: 11-14 years	Urine of children (9-12)
13 years			Physical growth during elementary school
Hokkaido cohort
4 months	Physical growth
1 year		ISAAC, ATS-DLD, infections	Physical growth
1.5 years	KIDS, M-CHAT
2 years		ISAAC, infections	Physical growth
3 years	KIDS, SDQ
4 years		ISAAC, infections	Physical growth
5 years	SDQ, DCDQ
6 years	ADHD-RS, ASQ (SCQ)
7 years		ISAAC	Health check-up data	Home visit	Urine of children/house dust
8 years	ADHD-RS, Conners 3P		CBCL, WISC-IV
9-11 years		ISAAC, history of vaccination		Medical examination phase 1 (blood pressure, anthropometric examination, FeNO, ISAAC)	Peripheral blood/urine of children
12 years	SDQ/KIDSCREEN: 8-17 years		Tanner staging, onset of puberty	Event-related brain Potentials: 11-14 years	Urine of children/tap water
13 years			Physical growth during elementary school
16 years		ISAAC	Physical growth during junior high school	Medical examination phase 2 (Tanner staging, Blood pressure, anthropometric examination, Gripping power, 2D/4D, consumer-product-use, personal care products): 14-17 years	Peripheral blood/urine of children/house dust

2D/4D: 2nd and 4th digits ratios, ADHD-RS: Attention Deficit Hyperactivity Disorder-Rating Scale, ASQ: Autism Screening Questionnaire, ATS-DLD: American Thoracic Society-Division of Lung Disease, M-CHAT: Modified Checklist for Autism in Toddlers, BSID-II: Bayley Scales of Infant Development second edition, CBCL: Child Behavior Checklist, Conners 3P: Conner’s 3rd Edition for Parents, DCDQ: Developmental Coordination Disorder Questionnaire, DDST: Denver Developmental Screening Tests, EES: Evaluation of Environmental Stimulation, FeNO: fractional exhaled nitric oxide, FTII: Fagan Test of Infant Intelligence, ISAAC: International Study of Asthma and Allergies in Childhood, J-PSAI: Japanese Pre-School Activities Inventory, K-ABC: Kaufman Assessment Battery for Children, KIDS: Kinder Infant Development Scale, SCQ: Social Communication Questionnaire, SDQ: Strengths and Difficulties Questionnaire, WAIS-R: Wechsler Adult Intelligence Scale-Revised, WISC-III: Wechsler Intelligence Scale for Children third edition, WCST-KFS: Wisconsin Card Sorting Test (Keio Version), WISC-IV: Wechsler Intelligence Scale for Children fourth edition.

We have also measured biological markers as shown in Table 2. In the Sapporo cohort, we measured maternal fatty acids and triglyceride in maternal blood during pregnancy, as well as immunoglobulin (Ig) E, IgA, adipokines, steroid and reproductive hormones in cord blood. Thyroid-stimulating hormone (TSH) and free thyroxin (FT4) levels were obtained from Sapporo municipal screening data of mothers and infants. In the Hokkaido cohort, maternal folic acid and cotinine levels were measured during late pregnancy. Thyroid hormones and thyroid antibodies of maternal blood in early pregnancy were also examined. Adipokines and thyroid hormones were measured in cord blood as well.

Table 2.

Biological markers measured in the Sapporo cohort and Hokkaido cohort

Sapporo cohort
Maternal blood	Thyroid hormones (TSH, FT4) Fatty acids (palmitic, stearic, palmitoleic, oleic, linoleic, arachidonic, α-linolenic, EPA, DHA) Triglyceride
Cord blood	IgE, IgA Adipokines (adiponectin, leptin) Steroid hormones (estradiol, testosterone, progesterone, cortisol, cortisone, DHEA, androstenedione) Reproductive hormones (LH, FSH, SHBG, prolactin, inhibin B, INSL-3)
Hokkaido cohort
Maternal blood	Thyroid hormones (TSH, FT3, FT4) and anti-thyroid antibodies (TPOAb, TgAb) Folic acid Cotinine
Cord blood	Adipokines (adiponectin, leptin, TNF-α, IL-6)

DHA: docosahexaenoic acid, DHEA: dehydroepiandrosterone, EPA: eicosapentaenoic acid, FSH: follicle-stimulating hormone, FT3: free triiodothyronine, FT4: free thyroxine, IgA: immunoglobulin A, IgE: immunoglobulin E, IL-6: interleukin-6, INSL-3: insulin-like factor 3, LH: luteinizing hormone, SHBG: sex hormone-binding globulin, TgAb: thyroglobulin antibody, TNF-α: tumor necrosis factor alfa, TPOAb: thyroid peroxidase antibody, TSH: thyroid-stimulating hormone.

3. Exposure levels of environmental chemicals in children

We have been conducting highly sensitive environmental measurements of exposure to chemicals with relatively low-levels (background) in daily life, particularly polychlorinated dibenzo-p-dioxins (PCDDs), and polychlorinated dibenzofuran (PCDF). These congeners were particularly problematic in the Kanemi oil poisoning (Yusho disease) case. The toxic equivalents (TEQs) of dioxins (29 congeners) were calculated using the toxic equivalency factor set of the WHO.¹⁷⁾ Per- and polyfluoroalkyl substances (PFAS) are a group of very stable and durable chemicals with water and oil repellent properties. There are more than 4000 PFAS with different numbers of carbon chains, and perfluorooctanesulfonic acid (PFOS), and perfluorooctanoic acid (PFOA), with 8 carbon chains, are representative compounds. PFOS (Annex B in 2009),¹⁸⁾ PFOA (Annex A in 2019)¹⁹⁾ and perfluorohexanesulfonic acid (PFHxS) (Annex A in 2022)²⁰⁾ are regulated under the Stockholm Convention due to their high persistence. Phthalates are plasticizers for plastics, with di(2-ethylhexyl) phthalate (DEHP) having the largest production volume. Phthalates are additives to polyvinyl chloride products such as flooring and wallpaper, as well as daily products and personal care products such as synthetic leather, paints, and cosmetics. The use of six phthalates (DEHP, dibutyl phthalate, butylbenzyl phthalate, di-iso-nonyl phthalate, di-iso-decyl phthalate, and di-n-octyl phthalate) is restricted as a precautionary measure in Japan (Ministry of Health, Labour and Welfare [MHLW] No. 267, 2002 and No. 336, 2010),²¹⁾ the United States, and Europe to prevent negative effects on health from toys, childcare products, food containers and packaging, especially those for oily foods. The half-lives of phthalates are short, ranging from a few hours to a few days, so they are not persistent; however, because they are found in daily commodities, people are constantly and ubiquitously exposed to them. Bisphenol A (BPA) is a substance in raw materials such as polycarbonate resin and epoxy resin. In Japan, the Food Sanitation Act limits the dissolution test standard for BPA from polycarbonate utensils and containers/packaging to 2.5 μg/ml (2.5 ppm) or less. The industry has voluntarily regulated the leaching of BPA in sealants for canned foods. On the other hand, various bisphenols are used as substitutes for BPA and there is a movement in Europe to regulate their endocrine-disrupting effects.

Table 3 shows the concentrations of each class of chemicals of the Sapporo cohort and the Hokkaido cohort. In the Sapporo cohort, in cooperation with the Fukuoka Prefectural Institute of Environmental Health, we used high-precision gas chromatography-mass spectrometry for the first time in the world to determine the exposure concentrations of 66 dioxins and PCBs congeners. We also measured organochlorine pesticides, PFOS, PFOA, phthalate metabolite mono(2-ethylhexyl) phthalate (MEHP), and BPA. The target chemicals were detected in almost all maternal blood. Mercury was measured using maternal hair collected after delivery. In the Hokkaido cohort, we conducted measurements of 11 PFAS, 7 metabolites of phthalates, and BPA.^16),22)

Table 3.

Exposure levels of environmental chemicals in the Sapporo and Hokkaido cohorts

					Percentile
	n	DL (range)	＞ DL (％)	Min	25th	50th	75th	Max
Sapporo cohort
Maternal blood
Total dioxins (TEQ pg/g lipid)	426	n/a	n/a	3.17	9.95	13.9	18.2	43.4
Total PCBs (ng/g lipid)	426	n/a	n/a	17.8	73.0	107	148	41500
p,p′-DDE	379	0.60	100	99.5	401	651	1010	4580
PFOS (ng/mL)	447	0.5	100	1.30	3.40	5.20	7.00	16.2
PFOA (ng/mL)	447	0.5	92.8	0.25	0.80	1.30	1.80	5.30
MEHP (ng/mL)	493	0.278	100	1.94	5.82	9.95	16.3	102
BPA (ng/mL)	59	0.04	76.3	＜DL	0.040	0.057	0.072	0.419
Cord blood
BPA (ng/mL)	285	0.04	68.8	＜DL	＜DL	0.051	0.076	0.217
Maternal hair
Me-Hg (μg/g)	430	n/a	100	0.24	0.96	1.40	1.89	7.55
Hokkaido cohort
Maternal blood
PFHxS (ng/ml)	1985	0.2	81.2	＜DL	0.2	0.3	0.4	3.4
PFHxA (ng/ml)	1985	0.1	44.1	＜DL	＜DL	＜DL	0.1	0.7
PFHpA (ng/ml)	1985	0.1	31.5	＜DL	＜DL	＜DL	0.1	1
PFOS (ng/ml)	1985	0.3	100	0.8	2.6	3.4	4.7	17.9
PFOA (ng/ml)	1985	0.2	99.9	＜DL	1.3	2	3.3	24.9
PFNA (ng/ml)	1985	0.3	99.9	＜DL	0.9	1.2	1.6	13.2
PFDA (ng/ml)	1985	0.1	99.3	＜DL	0.4	0.5	0.7	2.4
PFUnDA (ng/ml)	1985	0.1	99.6	＜DL	1	1.4	1.9	5.9
PFDoDA (ng/ml)	1985	0.1	89.5	＜DL	0.1	0.2	0.2	0.7
PFTrDA (ng/ml)	1985	0.1	97.4	＜DL	0.2	0.3	0.4	1.3
MnBP (ng/ml)	2571	(0.44-0.90)	99.9	0.61	17.0	31.0	49.0	270
MiBP (ng/ml)	2571	(0.44-0.90)	96.9	0.22	2.60	4.30	7.30	36.0
MBzP (ng/ml)	2571	(0.19-0.46)	4.5	＜DL	＜DL	＜DL	＜DL	7.20
MEHP (ng/ml)	2571	(0.31-0.48)	91.3	＜DL	0.5	1.20	7.00	150.0
MEHHP (ng/ml)	2571	(0.11-0.23)	16.2	＜DL	＜DL	＜DL	＜DL	3.30
MECPP (ng/ml)	2571	(0.11-0.21)	85.8	＜DL	＜DL	0.24	0.35	12.0
cx-MiNP (ng/ml)	2571	(0.12-0.15)	0.2	＜DL	＜DL	＜DL	＜DL	0.73
BPA (ng/ml)	1454	(0.008-0.04)	85.2	＜DL	0.04	0.07	0.18	56.0

BPA: bisphenol A, cx-MiNP mono-carboxy-isononyl phthalate, DL: detection limit, MBzP: mono-benzyl phthalate, MECPP: mono (2-ethyl-5-carboxypentyl) phthalate, Me-Hg: methylmercury, MEHHP: mono (2-ethyl-5-hydroxyhexyl) phthalate, MEHP: mono(2-ethylhexyl)phthalate, MiBP: mono-isobutyl phthalate, MnBP: mono-n-butyl phthalate, PCB: polychlorinated biphenyl, PFDA: perfluorodecanoic acid, PFDoDA: perfluorododecanoic acid, PFHpA: perfluoroheptanoic acid, PFHxA: perfluorohexanoic acid, PFHxS: perfluorohexanesulfonic acid, PFNA: perfluorononanoic acid, PFOA: perfluorooctanoic acid, PFOS: perfluorooctanesulfonic acid, PFTrDA: perfluorotridecanoic acid, PFUnDA: perfluoroundecanoic acid, p,p′-DDE: 1-dichloro-2,2-bis(p-chlorophenyl) ethylene, TEQ: toxicity equivalency quantity. Sapporo cohort data,¹⁵⁾ Hokkaido cohort PFAS data,²²⁾ Hokkaido cohort phthalate metabolites and BPA [unpublished data].

Regarding exposure levels, the median maternal total dioxin concentration in blood from the Sapporo cohort was 13.9 TEQ pg/g lipid,¹⁵⁾ which was lower than that of previous studies in Japan, as shown in Table 4 (other studies in Sapporo: 10.0 TEQ pg/g lipid),²³⁾ Osaka: 24.5 TEQ pg/g lipid,²⁴⁾ and Tohoku: 16.7 TEQ pg/g lipid).²⁵⁾ Studies in the Netherlands (29.8 TEQ pg/lipid),²⁶⁾ Germany (26.2 TEQ pg/lipid),²⁶⁾ the United States (New York: 11.8 TEQ pg/lipid),²⁶⁾ China (8 TEQ pg/g lipid),²⁷⁾ Korea (5.5 TEQ pg/g lipid),²⁸⁾ and Taiwan (14.6 TEQ pg/g lipid)²⁹⁾ found levels comparable to or lower than in our Japan report. We first reported the linear transfer of PFOS to infants via the placenta (approximately 30％) in the Sapporo cohort in 2004.³⁰⁾ Furthermore, as shown in Table 5, the PFOS exposure level (5.2 ng/mL) of maternal blood in the Sapporo cohort¹⁵⁾ was lower than that in the United States (24.4 ng/mL),³¹⁾ Denmark (8.1 ng/mL),³²⁾ and Norway (5.5 ng/mL)³³⁾ but higher than in Canada (4.6 ng/mL)³⁴⁾ and the Netherlands (1.6 ng/mL).³⁵⁾ Among Asian countries, Taiwan had the highest concentrations (5.9 ng/mL),³⁶⁾ whereas Korea (2.9 ng/mL)³⁷⁾ and China (2.1 ng/mL)³⁸⁾ had relatively low concentrations. Although the Hokkaido study was launched after PFOS, PFOA, and PFHxS were listed in the POPs Convention, we were able to note the decline of PFOS and PFOA levels through the years in maternal blood collected from 2003 to 2011.³⁹⁾

Table 4.

Concentrations of dioxin in human milk from 1999-2005 (TEQ pg/g lipid)

Country	Sampling period	PCDDs/PCDFs			CoPCBs
		Min	Median	Max	Min	Median	Max	Reference
Netherlands	2001-2003	17.09	18.27	21.29	10.9	11.57	13.08	26
Italy	2001-2003	9.4	12.66	14.83	11.02	16.29	19.33	26
Germany	2001-2003	11.14	12.53	12.72	12.8	13.67	14.31	26
USA	2001-2003	6.22	7.18	8.14	3.69	4.61	5.52	26
Japan: Sapporo	2002-2005	3.9	6.1	22	1.1	3.9	19	23
Japan: Tohoku	2001-2003	2.4	9.8	25	4.5	6.9	45.6	25
Japan: Osaka	1999	10	15.3	26.3	5.8	9.2	213	24
China	2002	3.5	5.1	5.7	2.46	2.89	3.16	27
Korea	2002		3.3＊			2.2＊		28
Taiwan	2000-2001		14.3					29

CoPCB: coplanar polychlorinated bipheny, PCDD: polychlorinated dibenzo-p-dioxin, PCDF: polychlorinated dibenzofuran. ^＊Mean value.

Table 5.

Maternal/cord blood levels of PFOS and PFOA in different studies (ng/mL)

Country	Name of the study	Sampling period	Sample type	PFOS	PFOA	Reference
Canada	MIREC	2008-2011	Maternal blood	4.6	1.7	34
USA	Project VIVA	1999-2002	Maternal blood	24.4	5.3	31
Denmark	Odense child cohort	2010-2012	Maternal blood	8.07	1.6	32
Norway	MoBa	2003-2004	Maternal blood	5.5	1.1	33
Netherlands	LINK	2011-2013	Cord blood	1.6	0.8	35
Japan	Hokkaido	2003-2011	Maternal blood	3.4	2.0	22
	Sapporo	2002-2005	Maternal blood	5.2	1.3	15
China	Guangzhou birth cohort	2013	Cord blood	2.1	1.2	38
Korea	Seoul	2008-2009	Maternal blood	2.9	1.4	37
Taiwan	TBPS	2004-2005	Cord blood	5.9	1.8	36

PFOA: perfluorooctanoic acid, PFOS: perfluorooctanesulfonic acid. Values are mean, median, or geometric mean.

4. Summary of findings from the Hokkaido Study (Sapporo cohort and Hokkaido cohort)

We have previously published findings to date from birth to the age of 8 years. Table 6 summarizes the health outcomes that have been reported in peer-reviewed English-language journals as of July 2023.

Table 6.

Findings from the Sapporo cohort and Hokkaido cohort

Outcome	Maternal exposures	Cohort	Findings when maternal exposure was increased	Reference
Birth size	Dioxins	Sapporo	2,3,4,7,8-PeCDF: Birth weight ↓ (β＝－24.5 g)	40
	PFAS	Sapporo	PFOS: Birth weight ↓ (β＝－269.4 g)	41
		Sapporo	PFOA: Birth weight ↓ (β＝－197 g)	42
		Sapporo	PFOA: IGF2 methylation ↓ (β＝－0.73)	43
		Hokkaido	PFNA: Birth weight ↓ (β＝－96.2 g) × 10 levels of PFDA: Birth weight ↓ (β＝－72.2 g)	22
	PCBs, MeHg	Sapporo	No association	44
	DEHP	Sapporo	Birth weight ↓	45
	Smoking	Sapporo	Birth weight ↓	46
		Hokkaido	High prenatal cotinine: Birth weight ↓	47
Thyroid hormones	PFAS	Sapporo	PFOS: Lower maternal TSH ↑, Infant TSH ↑	49
		Hokkaido	PFAS: FT3, TSH, FT4 (boys/girls) ↓ / ↑ depending on thyroid antibodies	53
	OH-PCBs	Sapporo	FT4: ↑ in mothers and neonates, no TSH effect	50
	BPA	Sapporo	No association	54
	DEHP	Sapporo	No association	55
Steroid hormones	PFAS	Sapporo	PFOS: E2 ↑, T/E2 ↓, P4 ↓, inhibin B ↓ ; PFOA: → inhibin B ↑ in males	56
		Sapporo	PFOS: cortisol and cortisone ↓, DHEA ↑, PFOA: DHEA ↓	61
	Dioxin	Sapporo	Cortisol, DHEA, SHBG: (males/females) ↑ / ↓	57
	DEHP	Sapporo	Cortisol, cortisone: ↓, DHEA ↑	58
	BPA	Sapporo	Testosterone, estradiol, progesterone: ↑ in boys	55
	Persistent organochlorine and pesticides	Sapporo	testosterone, testosterone/A-dione ratio, SHBG, or prolactin ↑ / ↓ and DHEA, estradiol/testosterone ratio ↑	60
Cord adipokines	PFAS	Sapporo	PFOS: Adiponectin ↑ (β＝0.12)	42
	DEHP	Sapporo	Leptin ↓, Adiponectin ↑	66
	BPA	Sapporo	Leptin ↓ (β＝－0.06)	66
Neurodevelopment	PFAS	Sapporo	PFOA: Mental development ↓ in girls at 6 months	76
		Hokkaido	PFAS: ADHD ↓ at 8 years	77
	DEHP	Hokkaido	No association	54
	BPA	Sapporo	Internalizing problems ↑ at 42 months	76
Behavioral issues Allergy/infectious diseases	Socioeconomic Factors	Hokkaido	Pre-pregnancy BMI, primiparity: Behavioral problems ↑	95
	Dioxin	Sapporo	IgE ↓, increase wheezing symptoms up to 7 years	80
	PFAS	Sapporo	PFOA: IgE ↓ in females	82
		Hokkaido	PFAS: increased prevalence of allergic symptoms and prevalence of infections in children up to 7 years	83
	DEHP	Sapporo	IgE ↓, increased prevalence of wheeze, otitis media, and chickenpox	81

ADHD: attention deficit hyperactivity disorder, A-dione: androstenedione, BMI: body mass index, BPA: bisphenol A, DEHP: di(2-ethylhexyl) phthalate, DHEA: dehydroepiandrosterone, IgE: immunoglobulin E, FT3: free triiodothyronine, FT4: free thyroxine, IGF2: insulin-like growth factor 2, MeHg: methyl-mercury, OH-PCB:, hydroxy polychlorinated biphenyl, P4: progesterone, PCB: polychlorinated biphenyl, PeCDF: pentachlorodibenzofuran, PFAS: per- and polyfluoroalkyl substance, PFDA: perfluorodecanoic acid, PFNA: perfluorononanoic acid, PFOA: perfluorooctanoic acid, PFOS: perfluorooctanesulfonic acid, SHBG: sex hormone binding globulin, T/E2: testosterone/estradiol ratio, TSH: thyroid-stimulating hormone. β represents the change in outcomes associated with a 10-fold increase in the exposure chemical. ↓: decreased; ↑: increased.

4.1. Birth size.

Maternal factors such as smoking, caffeine intake, educational history, the mother’s pre-pregnancy body mass index (BMI), and assisted reproductive treatment were found to have significant negative impacts on birth weight in both the Sapporo cohort and Hokkaido cohort.^22),40)-47) Higher levels of dioxins and phthalates in maternal blood were related to significantly smaller birth weight in the Sapporo cohort.⁴⁰⁾ Among PFAS, higher levels of maternal PFOS were associated with lower birth weight, especially for girls, in the Sapporo cohort,^41)-43) whereas higher levels of perfluorononanoic acid (PFNA) and perfluorodecanoic acid (PFDA) were associated with lower birth weight in the Hokkaido cohort. Higher levels of maternal PFNA were also associated with reduced birth length in the Hokkaido cohort.²²⁾ No associations with birth size were observed for mercury, PCBs, or DEHP at the exposure levels of the Sapporo cohort.⁴⁵⁾ Recently, the developmental origin of health and disease theory posits that early life environment is critical for health at later age. Maternal smoking and high cotinine level were associated with reduced birth weight in both the Sapporo and Hokkaido cohorts.^46),47) Adverse health outcomes such as low birth weight, SGA, and preterm birth in particular have serious health effects not only during the neonatal period and infancy but also later in life.⁴⁸⁾ As shown in Table 6, not only maternal smoking but also passive smoking and exposures to dioxins, PFOS, PFOA, PFNA, PFDA, and DEHP reduced birth weight. Although the statistical models were adjusted for potential confounding factors such as maternal age, pre-pregnancy BMI, and parity, combined exposures to such chemicals have not yet been investigated. Thus, examining combined exposures is the next task.

4.2. Hormone levels at birth.

4.2.1. Thyroid hormones.

Dioxins, PCBs, hydroxy-PCBs, and organochlorine pesticides were positively associated with infant FT4 in the Sapporo cohort.^49)-52) In this cohort, higher levels of maternal PFOS showed an inverse association with maternal TSH, whereas there was a positive association with infant TSH, but no association with either mother or child FT4. In the Hokkaido cohort, 11 maternal PFAS did not show any association with maternal TSH, but PFOS was associated with infant TSH.⁵³⁾ PFHxS and PFNA were positively associated with maternal free triiodothyronine (FT3), whereas perfluoroundecanoic acid and PFDA were inversely associated with infant FT3 and FT4.⁵³⁾ PFAS were also inversely associated with maternal thyroglobulin antibody and infant thyroid peroxidase antibody in the Hokkaido cohort. Positive associations with maternal PFAS and infant TSH were found in both the Sapporo and Hokkaido cohorts. However, effects on PFAS and thyroid hormones were complicated, including modification effects of existing thyroid antibody. Phthalates and BPA showed no association with thyroid hormones in the Sapporo cohort.^54),55) Further studies with replicates as well as toxicological and biological evidence from other cohorts are warranted.

4.2.2. Reproductive hormones.

In the Sapporo cohort, seven steroid hormones (progesterone, testosterone, estradiol, dehydroepiandrosterone [DHEA], androstenedione, cortisol, and cortisone), luteinizing hormone, follicle-stimulating hormone, inhibin B, insulin-like factor-3 (INSL3), sex hormone binding protein (SHBG), and prolactin were analyzed.^56)-59) Higher maternal exposure to dioxins resulted in lower inhibin B levels in their offspring.⁵³⁾ Stratified analysis by sex showed that several compounds, including mirex, monachlor, dichloro-diphenyl-dichloroethylene, dieldrin, and toxaphene, were found in boys and girls who had a significant increase or decrease in testosterone, the testosterone/androstenedione ratio, SHBG or prolactin, and were positively correlated with DHEA and the estradiol/testosterone ratio.⁶⁰⁾ The PFOS concentration was negatively correlated with progesterone, inhibin B, INSL3, and the testosterone/estradiol ratio in boys (Fig. 2), and progesterone, sex hormone binding globulin, and prolactin were negatively correlated in girls.^15),61) In the combined analysis of males and females, a positive correlation was found between PFOS and DHEA, and there were negative correlations of PFOS with cortisol and cortisone (Fig. 2).⁶¹⁾ Negative correlations with MEHP and progesterone, the testosterone/estradiol ratio, inhibin B, and INSL3 were found in boys (Fig. 3).⁵⁹⁾ Positive correlations with DEHA and negative correlations with cortisol and cortisone were found in both sexes.⁵⁸⁾ Regarding BPA, the levels of BPA showed a positive correlation with testosterone and progesterone levels in cord blood among boys.⁵⁵⁾ Overall, we reported for the first time that prenatal exposure to environmental chemicals may disturb steroid hormone balance. Especially in boys, prenatal exposure to environmental chemicals such as PFAS and phthalates was shown to decrease levels of inhibin B and INSL3, which represent the function of Leydig and Sertoli cells, respectively, in the testis.^56),59) There was some evidence reported that PFAS inhibit steroid synthesis in the testis and testosterone release in Leydig cells in male mice.⁶²⁾ DEHP caused a decrease in the number of Leydig and Sertoli cells as well as in the concentration of testosterone in fetuses in animal studies.^63),64) Our findings suggested that prenatal exposures to PFOS and DEHP results in a reduction of testis development in humans similar to the findings in animal studies. In addition, as shown in Figs. 2 and 3, our results indicated that there are dose-response relationships for both PFOS and DEHP with inhibin B. There were significantly lower levels of inhibin B among the fourth quartile levels of PFOS (＞ 7.1 ng/mL) and MEHP (＞ 14.29 ng/mL) than in the first quartile. Thus, these levels may define the fourth quartile as the lowest observed adverse effect level in humans. According to German human biomonitoring-II values, 10 ng/mL is defined as the level that, when exceeded, indicates a relevant health risk among women of child-bearing age.⁶⁵⁾ Our findings are in line with the HBM-II alert. We are still following-up these children to examine the effects on puberty and to determine if these disruptions of hormone levels in umbilical cord blood at birth continue. In addition, dioxins, PFOS, PFOA, and DEHP showed inverse associations with inhibin B among boys. As a future study, combined exposures to such chemicals need to be investigated.

Fig. 2.

Associations between maternal blood PFOS and child steroid and reproductive hormone levels at birth. DHEA: dehydroepiandrosterone, PFOS: perfluorooctanesulfonic acid, Q: quartile, T/E2: testosterone/estradiol ratio. X-axis: PFOS quartiles, Y-axis: each hormone level. Center point represents the adjusted least square mean (95％ confidence intervals), back transformed from log₁₀ to normal values, and the error bars depict the upper and lower 95％ confidence intervals. PFOS levels were divided into 4 categories; 1.5 ≤ Q1 ≤ 3.9, 3.9 ＜ Q2 ≤ 5.2, 5.2 ＜ Q3 ≤ 7.1, 7.1 ＜ Q4 ≤ 16.2 ng/mL. Results are based on multiple linear regression models that were adjusted for maternal age, parity, body mass index before pregnancy, annual income, smoking during pregnancy, caffeine consumption during pregnancy, gestational weeks of blood sampling for PFOS, and gestational age at birth. ^＊＊p ＜ 0.01 compared with the Q1, as calculated using the Hsu-Dunnett method.^15),61)

Fig. 3.

Associations between maternal blood MEHP (DEHP metabolite) and inhibin B / INSL3 levels at birth. INSL3: insulin-like factor-3, DEHP: di(2-ethylhexyl) phthalate, MEHP: mono(2-ethylhexyl)phthalate. X-axis: MEHP quartiles, Y-axis: each hormone level. Center: points represent adjusted least square mean (95％ confidence intervals) back transformed from log₁₀ to normal values, and the error bars depict the upper and lower 95％ confidence intervals. MEHP levels were divided into 4 categories; 1st ≤ 6.36, 6.37 ≤ 2nd ＜ 10.25, 10.25 ≤ 3rd ＜ 14.28, 14.29 ng/mL ≤ 4th. Statistical significance of the p value was ^＊＊p ＜ 0.002 based on the Bonferroni correction. Least-squares means were adjusted for maternal age, smoking during pregnancy, alcohol consumption during pregnancy, gestational age, and the blood sampling week.⁵⁹⁾

4.3. Adipokines.

Adipokines, which are bioactive substances secreted from adipocytes, have been reported to affect the relationship between leptin and adiponectin concentrations in cord blood and birth weight in previous epidemiological studies. In particular, leptin is known to be produced and secreted in the placenta and is considered important for fetal growth and development. We examined whether adiponectin and leptin concentrations in cord blood were related to environmental chemical concentrations in maternal blood. In the Sapporo cohort, an increase in the concentration of PFOS in maternal blood increased the concentration of adiponectin in cord blood.⁴²⁾ In the Hokkaido cohort, it was found that the concentration of leptin in cord blood decreased when the concentrations of BPA, DEHP, and di-n-butyl phthalate in maternal blood increased.⁶⁶⁾ Overall, prenatal PFOS and DEHP increased adiponectin, whereas DEHP and BPA reduced the leptin level at birth. In adults, adiponectin and leptin are indicators of obesity and metabolic syndrome. Thus, it is necessary to carefully monitor adipocytokine levels in children during the growth process.

4.4. Pre-school activities inventory and second/fourth finger digit ratio.

In relation to hormone balance alteration, it has been reported that higher fetal exposures to dioxins, PCBs, and phthalates are associated with a lower play score for boys in infancy,^67),68) suggesting that reduced androgen exposure in the fetal period affects brain sex differentiation. Therefore, in the Hokkaido Study, the second/fourth finger ratio was measured using photocopies of schoolchildren’s palms as surrogate markers of prenatal androgens. In the Sapporo cohort, the results showed that when the INSL3 concentration was low in the fetal period, the second/fourth finger ratio in school-age boys was greater, suggesting a feminine tendency.⁶⁹⁾ Furthermore, in the Hokkaido cohort, when the second/fourth finger ratio was greater in boys, their play scores showed a less male-typical play behavior tendency at school age.⁷⁰⁾ Further analysis of the effects of prenatal exposure to environmental chemicals on the second/fourth finger ratio and play tendencies should be continued.

4.5. Neurodevelopment.

Regarding neurodevelopment, we will first introduce our findings from the Sapporo cohort. In this cohort, we conducted face-to-face examinations to investigate children’s neurodevelopment using the Bayley Scale of Infant & Toddler Development-II (BSID-II) at 6 and 18 months of age (Table 7). The mental developmental index (MDI) and psychomotor developmental index (PDI) showed score differences for infants aged 6 and 18 months in relation to the 10-fold increased level of dioxins in maternal blood adjusted for gestational age, maternal smoking, and blood sampling time. Only PCDD congener 1,2,3,4,6,7,8-heptachlorodibenzo-p-dioxin was significantly negatively associated with MDI. PCDD congeners (1,2,3,7,8,9-hexachlorodibenzo-p-dioxin and 1,2,3,4,5,7,8-heptachlorodibenzo-p-dioxin), and PCDF congeners (2,3,7,8-tetrachlorodibenzofuran, 1,2,3,7,8-pentachlorodibenzofuran, 1,2,3,6,7,8-hexachlorodibenzofuran (HxCDF), and 2,3,4,5,6,7,8-HxCDF) were significantly negatively associated with PDI.⁷¹⁾ We also found inverse associations with in utero exposure to a dioxin and PCB mixture and infant PDI.⁷²⁾ Notably, the effect on BSID-II at 18 months was smaller than at 6 months, suggesting that the effects of fetal exposure became less detectable with growth.⁷³⁾ In terms of prenatal organochlorine pesticide exposure and neurodevelopment, our data showed no association with BSID-II scores at 6 months of age, but a significant negative association between cis-heptachlor epoxide and mental development scores at 18 months.⁷⁴⁾ Measures of intellectual functioning determined using an individualized intelligence test (Kaufman Assessment Battery for Children) at 42 months showed no overall negative association with dioxin-like compounds (DLCs). However, when analyzed by gender, there was a significant positive association between the knowledge and skill acquisition scale and DLCs in girls and a significant negative association between the cognitive processing process scale and DLCs in boys, suggesting that the direction of influence differed by sex.75) The relationship between BPA and neurodevelopment, the BSID-II at 6 and 18 months and the Child Behavior Checklist at 42 months of age were examined and a significant association was found only with problematic behavior at 42 months.⁷⁶⁾

Table 7.

Multiple linear regression analysis for effects dioxin-like compounds in maternal blood using MDI and PDI scores for 6- and 18-month-old children in the Sapporo cohort

	6 M MDI		6 M PDI		18 M MDI		18 M PDI
	βa	t	βa	t	βa	t	βa	t
Male
Total PCDD	－0.10	－0.91	－0.24	－2.36＊	0.01	0.12	－0.13	－0.89
Total PCDF	－0.05	－0.41	－0.18	－1.77	－0.13	－0.93	－0.15	－0.99
Total PCDD/PCDF	－0.10	－0.91	－0.24	－2.37＊	0.01	0.08	－0.13	－0.90
Total non-ortho PCBs	0.00	0.01	－0.20	－1.96＊	－0.13	－0.98	－0.02	－0.12
Total mono-ortho PCBs	－0.05	－0.48	－0.21	－2.16＊	－0.10	－0.79	0.03	0.18
Total coplanar PCBs	－0.05	－0.48	－0.21	－2.16＊	－0.10	－0.79	0.03	0.17
Total dioxin	－0.05	－0.49	－0.22	－2.23＊	－0.10	－0.79	0.03	0.18
Female
Total PCDD	－0.13	－1.18	－0.13	－1.19	0.20	1.34	0.00	0.02
Total PCDF	0.08	0.69	－0.12	－0.99	0.24	1.55	－0.10	－0.59
Total PCDD/PCDF	－0.13	－1.14	－0.13	－1.20	0.20	1.36	0.00	－0.01
Total non-ortho PCBs	0.04	0.35	－0.14	－1.29	0.15	1.02	0.03	0.22
Total mono-ortho PCBs	0.05	0.44	－0.16	－1.50	0.23	1.67	0.03	0.19
Total coplanar PCB	0.05	0.44	－0.16	－1.50	0.22	1.67	0.03	0.19
Total dioxin	0.04	0.38	－0.16	－1.53	0.23	1.69	0.02	0.16

MDI: Mental Development Index, PCB: polychlorinated biphenyl, PCDD: polychlorinated dibenzo-p-dioxin, PCDF: polychlorinated dibenzofuran, PDI: Psychomotor Development Index. ^＊p ＜ 0.05. β^a is the point increase in the developmental score per PCB and dioxin level (common logarithm) adjusted for gestational age (days), caffeine intake during pregnancy (mg/day), economic status; annual income, index of childcare environment, smoking during pregnancy, blood sampling time and fish intake during pregnancy (inshore fish and deep-sea fish).⁷¹⁾

For PFOA, associations with BSID-II were found in girls at 6 months of age, but no differences were seen at 18 months or among boys in the Sapporo cohort.⁷⁷⁾ In the Hokkaido cohort, associations between 11 PFAS and ADHD at 8 years of age were examined, and a rather low risk association was found.⁷⁸⁾ Overall, the association between prenatal PFAS exposure and child neurodevelopment was not clear. Interpretation of the protective effects of PFAS and ADHD should be done with caution, and further studies are needed to clarify this issue.

4.6. Immune functions.

In the Sapporo cohort, higher maternal dioxin levels were significantly associated with lower cord blood IgE levels and an increased risk of otitis media, the most frequent infectious disease up to 18 months of age. Relatively high levels of PCDFs were associated with a significantly increased risk for otitis media, although this association was only found in boys.⁷⁹⁾ This result was consistent with the results reported for high-concentration exposures, such as with Yusho. This was the first report in the world that even low-concentration exposure in daily life has effects on the immune system in children.⁷⁹⁾ Postnatal follow-up found no obvious effects at 3.5 years of age, but at 7 years of age, a 7.8-fold increased risk of wheezing in children due to prenatal dioxin exposure was observed, indicating a persistent effect on the immune system.⁸⁰⁾ The DEHP levels of mothers during the second to third trimester were related to lower cord blood IgE and a lower risk of wheezing at age 3.5 years, but increased risks of food allergy, otitis media and chicken pox at the age of 7 years.⁸¹⁾

For PFOA, prenatal exposure was associated with lower cord blood IgE in the Sapporo cohort.⁸²⁾ Moreover, we further examined the associations between PFAS and immune function by following the children in the Hokkaido cohort. Although there was a lower risk of allergic symptoms, an increased risk of the prevalence of infections was found, and this association continued at 4 and 7 years of age, suggesting long-term effects of immune suppression.^82)-84) Thus, further research into the relations between PFAS and antibodies in the Japanese population is warranted.

4.7. Genetic susceptibility predisposition.

To investigate genetic susceptibility, we examined the genotypes of aryl hydrocarbon receptor (AHR), cytochrome P450 1A1 (CYP1A1), and glutathione S-transferase mu 1, NAD(P)H quinone dehydrogenase 1,⁸⁵⁾ and cytochrome P450 2E1,⁴⁶⁾ in mothers and the X-ray repair cross-complementing protein 1 involved in DNA repair protein 1 gene polymorphism.⁴⁶⁾ Offspring of smoking mothers with certain genotypes of these polymorphisms had lower birth weight depending on the combination of the polymorphisms. The effect of passive smoking was examined by measuring the cotinine concentration in maternal plasma with high sensitivity (limit of detection: 0.12 ng/mL) and classifying the mothers in the Hokkaido cohort into three groups from the low concentration side: those having no exposure to smoking, exposure to passive smoking, and active smoking. The dose-dependent smoking exposure associated with birth size differed significantly depending on AHR and X-ray repair cross-complementing protein 1 gene polymorphisms.⁸⁶⁾ Furthermore, we found that maternal passive smoking during pregnancy and CYP1A1 gene polymorphisms influenced the association between head circumference development of the infants up to 3 years of age.⁸⁷⁾ The concentration of dioxins in maternal blood varied according to the mother’s CYP1A1 gene polymorphism, which is involved in dioxin metabolism; among the GG, GA, and AA genotypes of AHR and the dominant genotype model [TT ＋ TC] vs. CC of CYP1A1 (rs4646903)⁸⁸⁾ in the Sapporo cohort. Furthermore, the association between maternal blood dioxin concentrations and birth weight differed significantly depending on the glutathione S-transferase mu 1 gene polymorphism of the mother.⁸⁹⁾ To the best of our knowledge, there has been no study that examined dioxins and AHR other than ours. Dioxin activates AHR, so the gene-environment interaction of dioxin and AHR with birth size needs to be investigated in future studies. Overall, there is some evidence of the effect of gene-environment interactions on lower birth weight. This knowledge on gene-environment interactions suggests the necessity of future studies on the combined effects of maternal-child genetic polymorphisms and exposure to environmental chemicals.

4.8. Epigenetic modification.

Regarding the modificatory effects of genetic DNA (epigenomic analysis), which have attracted great attention in recent years, we reported DNA methylation by PFAS, phthalates, BPA, and DEHP in a comprehensive analysis of 450,000 locations.^90)-92) We also measured cord blood DNA methylation frequencies of long interspersed nuclear element-1 (LINE-1), a marker of genome-wide DNA methylation, insulin-like growth factor 2 (IGF2), and H19 regions, which play important roles in fetal development. We found a significant negative association between PFOA exposure and IGF2 methylation.⁴³⁾ Furthermore, IGF2 methylation mediated a decrease in the ponderal index, explaining 20％ of the effect of PFOA exposure. Thus, it is suggested that PFAS exposure affects fetal growth through its effect on DNA methylation.⁴³⁾ In addition, a positive association between decachlorobiphenyl congener concentrations and H19 methylation, and a positive association between heptachlorobiphenyl congeners and LINE-1 methylation were observed in fetal exposure to PCBs. After stratification by sex, these effects were found in girls.⁹³⁾

4.9. Impact of socioeconomic factors.

To examine parental factors, we used three definitions of birth outcomes: preterm birth (＜ 37 weeks), very low birth weight (＜1,500 g), and term-SGA (after 37 weeks). Our findings showed that when the parents had a university degree or postgraduate degree there was a reduced risk of SGA. Regarding the influence of socioeconomic factors, an interaction was observed between maternal education and BMI for the risk of term-SGA, indicating that the risk may be highest for mothers whose last education was a secondary school degree and whose BMI was less than 18.5.⁹⁴⁾ An analysis of more than 2,000 participants in the Hokkaido cohort also revealed that children’s problem behaviors at age 5 (Strengths and Difficulties Questionnaire) were significantly associated with a maternal pre-pregnancy BMI of 30 or higher (obesity), maternal educational background of less than high school graduation, annual household income of less than 3 million yen during pregnancy, first child, and the child being a boy.⁹⁵⁾ The findings suggested the need for school education from lower grades about the health effects of actions such as tobacco smoking and improving food intake. Moreover, the effects of chemicals in the environment on children’s health, should also be carefully considered in reference to socioeconomic factors.

5. Future issues and suggestions

This report presents part of the findings of the Hokkaido Study, the first birth cohort study in Japan. The most important point to emphasize is that although PCBs, dioxins, and organochlorine chemicals have been regulated by the Stockholm Convention and their concentrations in the environment are decreasing worldwide, certain levels still remain in our blood due to their long half-lives. Even though the concentrations in the Sapporo and Hokkaido cohorts in Japan were relatively low, effects of exposure during the fetal period were associated with various outcomes. One reason for this is the development of improved instruments and measurement techniques. The measurement of exposure levels for more than 2,500 persons in this study revealed health effects at low levels that were previously invisible. In addition, PFAS such as PFOS and PFOA have been detected in recent years, and our data indicated that various effects of PFAS can be observed in Japan. Chemicals with short half-lives, such as phthalates and BPA, show endocrine-disrupting effects not only in animal studies but also on human sex hormone levels in cord blood. Furthermore, the exposure levels of phthalates in 7-year-old children have been stable, whereas the exposure levels of 1,2-cyclohexane dicarboxylic acid diisononyl ester and di-2-ethylhexyl terephthalate, which are alternatives to phthalates, have been increasing, indicating the need to evaluate the health effects of these substitutes.

In the future, we will continue to deal with the following issues: (1) How long the effects of outcomes seen in childhood continue and if the effects become more or less detectable. To answer these questions, long-term follow-up is very important. Especially for developmental disorders, we need to take into account the living environment after birth and clarify details such as how long the effects of prenatal exposure persist or if there is any catch-up due to the postnatal environment. All of these findings will provide valuable data that can be easily understood by the general public. (2) Regarding causality, it is necessary to further examine whether the effects on outcomes are independent or combined effects of individual chemicals. (3) For both exposure and outcome assessment, it is important to work with countries around the world to investigate the hazardous environments in each region and to determine the magnitude of risk of next-generation effects from environmental chemicals in order to create preventive measures. (4) We have already conducted a risk assessment of the effects of low-concentration exposure on the next generation in our epidemiological study in Japan, and many findings are being obtained. These findings should be reflected in policy. For example, the results of risk assessment of PFAS with long carbon chains such as PFNA and PFDA, the exposure levels of which are increasing in Japan, can be utilized in regulations and other measures, especially compared with PFOS and PFOA, the concentrations of which have decreased under the Stockholm Convention. (5) We have accumulated a great deal of data in our study initiated in 2001 under the aegis of the Japanese MHLW, 10 years prior to the beginning of the large-scale Japan Environment and Children’s Study supported by the Ministry of the Environment. To make good use of the national taxpayers’ money, we would like to emphasize the importance of the results of the Hokkaido Study for the implementation of national health policy and environmental measures.

As a future perspective for extending the Hokkaido study, we envision continuing to follow the participating children into adulthood and aspire to establish a next-generation cohort study. Furthermore, a key limitation of our current research is its observational design, which limits our ability to reach causal conclusions and determine dose-response relationships. To address these gaps, we propose integrating epidemiological approaches with experimental research as well as investigating combined exposures as suggested by Tohyama and Honda.⁹⁶⁾ Such interdisciplinary collaboration efforts will provide valuable understanding of the mechanisms in the associations and enhance the significance of the Hokkaido Study findings.

In 2011, the Birth Cohort Consortium of Asia was established by the principal investigators of cohorts in Japan (the Hokkaido Study), Korea (MOCHE), and Taiwan (TBPS).⁹⁷⁾ The results of our epigenomic analyses are the scientific outcome of the collaboration among these Birth Cohort Consortium of Asia cohorts.^90),91) Moreover, we have collaborated with young PhD candidates using the Sapporo Cohort data investigating the effects of mixtures of dioxins on neurodevelopment.⁷²⁾ The availability of these data and collaboration with the Hokkaido Study will contribute to the development of young researchers who wish to gain experience in environmental epidemiology. In addition, the Center for Environmental and Health Sciences at Hokkaido University, to which the authors belong, was designated a WHO Collaborating Centre (WHOCC; JPN-91) for Environmental Health and Prevention of Chemical Hazards in April 2015. The Center is currently in its third term of operation as a WHOCC. It is expected to further develop joint research with Asian countries and apply the results to chemical management and the prevention of health hazards in the future. Moreover, the effects of various environmental chemical exposures on children, including the sources of such exposures related to the indoor environment, can reviewed using our recently published books as further references.^98)-100) As part of the contribution to the WHOCC, we have created and published informative leaflets on dioxins,¹⁰¹⁾ PFAS,¹⁰²⁾ pesticides,¹⁰³⁾ and lead,¹⁰⁴⁾ outlining their levels, health effects and management recommendations with a focus on the Western Pacific Region.

Conflict of interest

The authors have no conflicts of interest to declare.

Acknowledgements

We would like to express our appreciation to all of the study participants of the Hokkaido Study on Environment and Children’ Health. We also express our profound gratitude to all personnel in the hospitals and clinics who collaborated with the study: Sapporo Toho Hospital, Aoba Ladies Clinic, Akiyama Memorial Hospital, Asahikawa Medical College Hospital, Asahikawa Red Cross Hospital, Iwamizawa Children’s Clinic, Engaru-Kosei General Hospital, Endo Kikyo Maternity Clinic, Ohji General Hospital, Obihiro-Kyokai Hospital, Obihiro-Kosei General Hospital, Kitami Red Cross Hospital, Kitami Lady’s Clinic, Kin-ikyo Sapporo Hospital, Kushiro Red Cross Hospital, Kushiro Rosai Hospital, Keiai Hospital, Kohnan Hospital, Memuro Municipal Hospital, Gorinbashi Hospital, Sapporo Medical University Hospital, Sapporo-Kosei General Hospital, Sapporo Tokushukai Hospital, Sapporo City General Hospital, Shibetsu City General Hospital, Hakodate City Hospital, Wakkanai City Hospital, Shiroishi Hospital, Steel Memorial Muroran Hospital, Teine Keijinkai Hospital, Tenshi Hospital, Hokkaido Monbetsu Hospital, Nakashibetsu Municipal Hospital, Nakamura Hospital, Nayoro City General Hospital, Nikko Memorial Hospital, Hakodate Goryoukaku Hospital, Hakodate Central General Hospital, Hashimoto Clinic, Hoyukai Sapporo Hospital, Hokkaido University Hospital, Kushiro City General Hospital, Iwamisawa Genecology and Pediatric Clinic, Steel Memorial Muroran Hospital, and Teine Keijinkai Hospital.

This study is being conducted in cooperation with the Hokkaido University Center for Environmental and Health Sciences, the Department of Public Health, Hokkaido University Graduate School of Medicine, the Department of Reproductive and Developmental Medicine, Hokkaido University Graduate School of Medicine, the Department of Pediatrics, Hokkaido University Graduate School of Medicine, the Department of Obstetrics and Gynecology, Sapporo Medical University, the Department of Obstetrics and Gynecology, Asahikawa Medical University and Toxicology Laboratory, and the Faculty of Veterinary Medicine, Hokkaido University. The study also conducts joint research with the Fukuoka Institute of Health and Environmental Sciences, the Department of Analytical Chemistry, Hoshi University, the National Institute for Minamata Disease and the Sapporo City Institute of Public Health.

The study is funded by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) and the Ministry of Health, Labour and Welfare (MHLW) of Japan.

Non-standard abbreviation list

ADHD

attention deficit hyperactivity disorder

AHR

aryl hydrocarbon receptor

BMI

body mass index

BPA

bisphenol A

BSID-II

Bayley Scale of Infant & Toddler Development-II

CYP1A1

cytochrome P450 1A1

DEHP

di(2-ethylhexyl) phthalate

DHEA

dehydroepiandrosterone

DLC

dioxin-like compound

FT3

free triiodothyronine

FT4

free thyroxine

Ig

immunoglobulin

IGF2

insulin-like growth factor 2

INSL3

insulin-like factor-3

MDI

mental developmental index

MEHP

phthalate metabolite mono(2-ethylhexyl) phthalate

MHLW

Ministry of Health, Labour and Welfare

PCB

polychlorinated biphenyl

PCDD

polychlorinated dibenzo-p-dioxin

PCDF

polychlorinated dibenzofuran

PDI

psychomotor developmental index

PFAS

per- and polyfluoroalkyl substance

PFDA

perfluorodecanoic acid

PFHxS

perfluorohexanesulfonic acid

PFNA

perfluorononanoic acid

PFOA

perfluorooctanoic acid

PFOS

perfluorooctanesulfonic acid

POPs

persistent organic pollutants

SGA

small for gestational age

TEQ

toxic equivalent

TSH

thyroid-stimulating hormone

WHO

World Health Organization

WHOCC

World Health Organization Collaborating Centre

Profile

Reiko Kishi, MD, born in 1947, is a distinguished researcher and professor known for her pioneering work in environmental health. Inspired by her studies on Minamata disease victims, she developed a strong interest in toxicology during her medical education. After earning her PhD from Hokkaido University in 1977 and an MPH from Harvard University in 1990, she became the first female full professor of Public Health Sciences at Hokkaido University in 1997. She currently serves as a Distinguished Professor at the university's Center for Environmental and Health Sciences.

Prof. Kishi's groundbreaking research focuses on the health impacts of low-level chemical exposures during early development. In 2002, she launched the Hokkaido Study of Environment and Children's Health, which examined the effects of background chemical exposure on children's health. Her team has discovered that substances such as per- and polyfluoroalkyl substances can influence birth weight, immune function, and hormone levels in cord blood. They also identified genetic factors that affect fetal vulnerability to these chemicals.

In 2013, Prof. Kishi was elected as a member of the Collegium Ramazzini, one of only 180 internationally recognized experts who belong to this academy dedicated to the improvement of occupation and environmental health and who have embraced the habit of truth.

In 2021, Prof. Kishi was honored with the prestigious John Goldsmith Award by the International Society for Environmental Epidemiology (ISEE), a lifetime achievement award recognizing her pioneering and dedicated academic work and outstanding research contributions. She became the first award recipient from outside Europe and North America in the field of environmental epidemiology.

Beyond her research, Prof. Kishi has contributed to shaping public health policies. As a member of the Science Council of Japan, she led efforts to prevent death from overwork (karoshi) and improve mental health in the workplace. After the Fukushima disaster in 2011, she helped establish radiation safety limits for food. She co-founded the Birth Cohort Consortium of Asia in 2011 and directs a WHO Collaborating Centre. Her work continues to drive science-based policies for achieving sustainable development goals.