Cohort Profile: The Mother-Child Cohort in Crete, Greece (Rhea Study)

Epidemiological and animal research shows that disease aetiology often has to be evaluated with a life course perspective, starting as close to the time of conception as possible and, furthermore, that genetic predisposition, environmental exposures and social context have to be considered simultaneously.1 Exposures during the prenatal and postnatal periods have implications for children’s health and may also have implications for morbidity occurring later in life, including obesity and cardiovascular diseases, asthma and allergy, cancer, delayed neurodevelopment, mental illnesses and more.2 Birth cohort research has a long history in Europe and extensive collaborations have been established.3 The Rhea cohort is the only pregnancy cohort established and followed up in Greece and one of the very few cohorts in a large area in southern Europe and the eastern Mediterranean where birth cohort research has just started to emerge, and some of the circumstances observed in the Rhea population actually reflect patterns in this wider region. The research team of Rhea study specializes on specific research areas, including obesity, neurodevelopment and child asthma, but also addresses social, environmental and lifestyle factors specific to the Greek society such as use of pesticides, adherence to the Mediterranean diet or psychosocial factors at a time of economic crisis. The name of the cohort ‘RHEA’ is inspired by ‘Ρέα’ the mythological wife of Kronos (Cronus). The Rhea study was established in 2007 in the island of Crete, aiming to recruit a representative sample of mother-child pairs from early pregnancy and to follow them up to young adulthood. The uniqueness of this cohort is that it takes place in the island of Crete where the beneficial effect of the Mediterranean diet was first described during the Seven Countries Study of Cardiovascular Diseases, started at the end of the 1950s. The typical Cretan Mediterranean diet was associated with lower incidence and mortality from cardiovascular diseases and also with the lowest death rates and the greatest survival rates.4 This has changed dramatically over the last 50 years, and the Rhea cohort was created in response to the rising concerns about the high rates of childhood obesity in Greece (more than 40% overweight and obese children at school age according to recent estimates).5 The general aims of the project are: to characterize nutritional, environmental and psychosocial determinants of children’s growth and development; to focus on four primary outcome areas of research: (i) offspring growth and obesity; (ii) neuropsychological and behavioural development; (iii) allergies and asthma in childhood; and (iv) genotoxicity; to evaluate mother’s health during and after pregnancy; and to evaluate the interaction between environmental stressors and genetic variants in children’s growth and health. The Rhea study has been funded by various European grants since 2006 and by the Greek Ministry of Health (see Funding section). The funding has been continuous and has allowed the study to accomplish multiple follow-ups and to produce research and scientific results. Pregnant women (Greek and immigrants) were recruited at the time of the first comprehensive ultrasound examination at around week 12 of gestation, from four antenatal clinics (two public and two private) in Heraklion city, covering the wider Heraklion region, during the 12-month period from February 2007 until February 2008. The inclusion criteria for study participants were: residents in the study area; pregnant women older than 16 years; first antenatal visit: hospitals or private clinics in Heraklion district; no communication handicap. Participants were invited to provide blood and urine samples and to participate in a face-to-face interview. The flow diagram of Rhea cohort participation is described in Figure 1. During the study period, 1610 pregnant women agreed to participate and 1363 singleton pregnancies were followed up until delivery. Late recruitment at birth included additional 95 mother-child pairs. Table 1 describes characteristics of the Rhea study mothers at recruitment and delivery, and of their children at each postnatal visit. Characteristics of mother-child pairs included in the RHEA cohort study Characteristics of mother-child pairs included in the RHEA cohort study Rhea study flow diagram. We initially contacted pregnant women in the first trimester of pregnancy at the time of the first routinely scheduled major ultrasound test (median 12 weeks of gestation). After the clinical visit was completed, specially trained midwives met privately with interested women to describe the study in greater detail, then obtained written informed consent and completed in person a detailed questionnaire on diet, environmental exposures and socio-demographic and lifestyle characteristics. Midwives measured height, weight and blood pressure and collected blood and urine samples from pregnant women. We next met pregnant women in the third trimester of pregnancy (median 32 weeks of gestation) when midwives administered questionnaires on maternal stress and collected urine samples and medical record information on pregnancy-related complications as well as fetal ultrasound measurements. During the birth admission (median 38 weeks of gestation) midwives measured the infant’s size (weight, length, anogenital distance and head circumference) and collected cord blood samples, as well as information on the type of delivery, infant sex and gestational age. At 8–10 weeks after giving birth, mothers were approached by phone and provided information about symptoms of postnatal depression. Subsequently, we completed in-person clinical visits with mothers and children during infancy (median ages 9 and 18 months), early childhood (median 4.2 years) and mid-childhood (median 6.5 years). Briefly, visits included a detailed child clinical examination (anthropometry, blood pressure measurement, screening for atopic dermatitis), spirometry, neurodevelopmental assessment and biological sample collection (blood, urine and hair samples). Figure 2 summarizes the visit frequency and timing. The majority of clinical visits were conducted at the University Hospital of Heraklion in Greece, but for families living in rural areas of Crete, we also organized visits at four Rural Health Services in Crete. At each visit, we obtained written informed consent from the mothers and the Ethics Committee of the University Hospital at Heraklion approved the study protocols. Schedule of study visits and summary of measures at each visit. Major efforts have been made by the study staff to keep the families involved in the study and minimize losses to follow-up. Specific strategies implemented to retain the study sample are described below: Personal invitations were sent to all participants by post before each follow-up, including informative bulletins to keep the families updated with the study. Personal feedback related to child’s growth measurements was given to the families after the clinical examination. In cases of overweight/obese children or with signs of neurodevelopmental delay, parents were advised to have an additional appointment with an expert. During the 4th and 6th year follow-ups, children who provided biological samples received feedback concerning their blood measurements. Research assistants focus on having a more personal communication with the families, responding promptly to questions or problems. Of the 1458 children (singleton births) followed up until delivery, including the later-recruited mother-child pairs at birth, 879 children completed the early childhood follow-up and 607 children the mid-childhood follow-up. Children who attended the follow-up tended to have older parents who were married, of higher education and of Greek origin. Attrition at follow-up was mainly due to withdrawal, difficulty in keeping track of changed addresses and telephone numbers, emigration, children’s severe illness and mothers’ unwillingness to attend the in-person visits (Figure 1). In most analyses, in order to account for missing data we used chained equations to multiply impute missing values. Tables 2 and 3 provide an overview of exposure and outcome data collected at different time points during the study. Exposure variables assessed via questionnaires include parental socio-demographic and lifestyle characteristics (smoking habits, sleep patterns, sedentary behaviours, physical activity, nursery attendance, hours spent with the child), pregnancy and delivery characteristics, household information (pets, moisture/moulds in house, bedroom location, cleaning products) and environmental factors (water use, pesticides use, sun exposure, noise, air pollution). Data collected on maternal and child exposures in the RHEA cohort study TSH, thyroid stimulating hormone; free T4, free thyroxine; IFNγ, interferon gamma; IL, interleukin, TNF α, tumour necrosis factor α; MIP1a, macrophage inflammatory protein 1-alpha; BPA, bisphenol A; OP pesticides, organophosphate pesticides; THMs, trihalomethanes; EBV, Epstein-Barr virus; CMV, cytomegalovirus; HSV-1, herpes simplex virus 1; HSV-2, herpes simplex virus 2; HHV-8, human herpesvirus 8; BKPyV, BK polyomavirus; JCPyV, JC polyomavirus; KIPyV, KI polyomavirus; WUPyV, WU polyomavirus; MCPyV, Merkel cell polyomavirus; TSPyV, trichodysplasia spinulosa polyomavirus; HPyV9, human polyomavirus 9; HPyV10, human polyomavirus 10. Data collected on maternal and child exposures in the RHEA cohort study TSH, thyroid stimulating hormone; free T4, free thyroxine; IFNγ, interferon gamma; IL, interleukin, TNF α, tumour necrosis factor α; MIP1a, macrophage inflammatory protein 1-alpha; BPA, bisphenol A; OP pesticides, organophosphate pesticides; THMs, trihalomethanes; EBV, Epstein-Barr virus; CMV, cytomegalovirus; HSV-1, herpes simplex virus 1; HSV-2, herpes simplex virus 2; HHV-8, human herpesvirus 8; BKPyV, BK polyomavirus; JCPyV, JC polyomavirus; KIPyV, KI polyomavirus; WUPyV, WU polyomavirus; MCPyV, Merkel cell polyomavirus; TSPyV, trichodysplasia spinulosa polyomavirus; HPyV9, human polyomavirus 9; HPyV10, human polyomavirus 10. Data collected on maternal and children’s health outcomes in the RHEA cohort study AGD, anogenital distance; BIA, bioelectric impedance analyses; ADHD, attention-deficit/hyperactivity disorder; SDQ, strengths and difficulties questionnaire; CBCL, child behaviour checklist. Data collected on maternal and children’s health outcomes in the RHEA cohort study AGD, anogenital distance; BIA, bioelectric impedance analyses; ADHD, attention-deficit/hyperactivity disorder; SDQ, strengths and difficulties questionnaire; CBCL, child behaviour checklist. Dietary habits of pregnant women have been assessed with the use of a validated food frequency questionnaire (FFQ), completed in mid-pregnancy (14th–18th weeks of gestation). We assessed infant feeding (breastfeeding, use of infant formula, age of introduction to solid foods) via an interview-administered questionnaire at 9 and 18 months. Dietary assessment at 4 and 6 years of age was performed with the use of a validated6 semi-quantitative FFQ. We also used the Children’s Eating Behaviour Questionnaire (CEBQ) to assess children’s eating behaviour, in a subsample of the Rhea study.7 We have extensive, multimodal and repeated assessments of environmental exposures in both mothers and children from parental questionnaires, biological samples and environmental measurements. To date, we have measured more than 100 different chemicals in collected biospecimens (blood, urine and hair samples). These include heavy metals, tobacco smoke metabolites, polychlorinated biphenyls (PCBs), dioxins, organochlorine pesticides [dichlorodiphenyl dichloroethene (DDE), dichlorodiphenyl trichloroethane (DDT) and hexachlorobenzene (HCB)], phthalate metabolites, environmental phenols, perfluoroalkyl substances, bisphenol A (BPA), parabens and organophosphate pesticide metabolites. For many chemicals, we have multiple measures during pregnancy and childhood. Chemicals in drinking water and outdoor air pollution (ESCAPE protocol) have been measured. The primary outcomes of the Rhea study are pregnancy and birth outcomes, offspring obesity and cardiometabolic risk factors, asthma and atopy, and neuropsychological and behavioural development. Fetal growth data including femoral length, biparietal diameter and head and abdominal circumferences were obtained during regular clinic visits the second and thirdrd trimesters by ultrasound measures. Child anthropometry measures included weight, length/ height and head circumference at birth, plus waist circumference and mid-upper arm and skinfold thickness at four sites of the body (subscapular, triceps, suprailiac and thigh) in infancy and early childhood (4- and 6-year follow-up) with very good intra-observer and inter-observer reliability. In mid-childhood (6-year follow-up) we additionally performed fat mass measurements by bioelectrical impedance analysis. Internal gender- and age–adjusted growth curves have been created within the cohort, which allow estimation of individual growth patterns. Anogenital distance (the distance from the anus to the genitalia) measurements at birth and early infancy (mean age17.5 ± 6.8 months) were performed in 897 children using Vernier digital callipers.8–10 We assessed wheeze and asthma symptoms by questionnaire in infancy and early and mid-childhood, and lung function via spirometry at the 4- and 6-year follow-ups. Novel asthma and allergy phenotypes have been defined based on the Mechanisms of the Development of ALLergy (MeDALL) protocol.11 Neurodevelopment has been an important outcome for many analyses in the Rhea cohort. Experienced psychologists carried out the first assessment at 18 months using the Bayley Scales of Infant and Toddler Development (Bayley-III).12 At 4 years, trained psychologists evaluated children’s abilities through the McCarthy Scales of Children Abilities13 and parents completed two other questionnaires regarding the presence and the severity of certain symptoms and behaviours of their children: the Attention-Deficit/Hyperactivity Disorder Test (ADHDT)14 and the Strengths and Difficulties Questionnaire (SDQ).15 Neurodevelopmental outcomes at 6 years were assessed through a battery of internationally standardized computer tests (N-BACK, Attention Network Test, Trail Making Test, Raven). Parents completed the Conners’16 and Child Behaviour Checklist (CBCL)17 questionnaires evaluating child’s behavioural and emotional problems. We assessed maternal mental health (anxiety, depressive symptoms and personality characteristics) during pregnancy and postpartum (8–10 weeks after delivery) with the use of self-administered questionnaires, and maternal cognition with the use of Raven’s Standard Progressive Matrices.18,19 Finally, we measured micronuclei (MN) frequencies in blood samples collected at pregnancy, birth and at 4 years of age. MN frequencies are biomarkers of early genetic effects that have been used to investigate the association between environmental exposures and cancer. We carried out the cytokinesis-block micronuclei (CBMN) assay using whole blood according to the standardized protocol for the semi-automated image analysis system.20–22 We recorded the total numbers of bi-nucleated T lymphocytes and calculated MN frequency per 1000 bi-nucleated T lymphocytes. We designed our biospecimen collection to identify unique windows of vulnerability to early life stressors during gestation or childhood (Table 2). Blood samples have been collected from pregnant women (n = 1400) and from children at birth (n = 900), at 4 years (n = 815) and at 6 years of age (n = 550). Urine samples are also available, collected during pregnancy (n = 1300), at 2 years (n = 390), at 4 years (n = 845) and at 6 years of age (n = 614). DNA has been extracted from 1300 pregnancy blood samples, 820 cord blood samples at birth and 670 blood samples at 4 years of age. Other biological samples include child saliva and hair samples collected at 2 years and 6 years of age. All samples have been processed within 2 h from sampling time, and aliquots of urine, whole blood, serum, plasma, red blood cells and buffy coat are currently stored in ultra-low-temperature freezers (−80°C). All data generated within the Rhea cohort, including those measured through laboratory assays, are identified by a unique ID code for each sample. Thus, personal details (e.g. name, surname, birth date) never appear on the samples and the generated data are untraceable. The Rhea project is still ongoing, with the analysis of data now in progress. The first paper based on Rhea data was published in 2009.23 The Rhea website gives updated references with a Pubmed link to all Rhea publications [http://www.rhea.gr/en/publications/]. A summary of published articles within the framework of the Rhea birth cohort study is given in Table 4. We also provide below a selection of key findings from the core topics of the Rhea study. Summary table of published articles of the RHEA cohort study DDE, dichlorodiphenyldichloroethene; HCB, polychlorinated gestational Summary table of published articles of the RHEA cohort study DDE, dichlorodiphenyldichloroethene; HCB, polychlorinated gestational We have that a high adherence to the Mediterranean diet during pregnancy was associated with and in lower risk of postpartum and lower and lower blood pressure in red during pregnancy was associated with micronuclei frequency in pregnant women and higher risk of child wheeze in the first year of Children who food at by processed and lower in general and of the Child eating behaviours such as food and emotional were associated with symptoms in exposure to HCB, and was associated with lower risk of childhood obesity, higher blood pressure and at 4 years of prenatal exposure to and was associated with anogenital distance in exposure during pregnancy was associated with children’s function measured at with in early pregnancy were at high risk for gestational and We associated with maternal by in early pregnancy, and that a of and were associated with both fetal growth and maternal a between maternal and birth was a of offspring obesity and and was associated with child in and early and behavioural and symptoms at 4 years of We measured and using in blood samples collected at birth and at 3 and years of age. We that early life social and environmental characteristics of and in to a virus was associated with symptoms at early childhood to was associated with lower maternal personality assessed during pregnancy, we that which to was associated with risk for fetal weight growth and fetal head circumference growth maternal and were associated with higher risk for depressive at weeks antenatal and postnatal of each associated with in offspring neurodevelopment at 18 months of The strengths of Rhea cohort are general population with the inclusion of women from different at early pregnancy, the in-person collection of a large of exposures and outcome the of repeated and the extensive assessment was conducted by trained and staff for the and third trimesters of and for the child 18 4 years, 6 years). We used standardized questionnaires, personal and extensive biological collection to the data All research staff were of the children’s exposures and health outcomes which to include the to follow-up over time, and of to study Attrition is an in most and is a as families be for follow-up. In the Rhea cohort, were more to have a lower with and this may have in the estimation of some efforts have been made to retain all mother-child pairs in the cohort, by the and personal and regular with the families by our research sample size allow the study of outcomes or for continuous data are with those of other To scientific we have in collaborations of birth cohort to study the of early life stressors on birth and child health The Rhea cohort is involved in many and collaborations and we with other interested in with the Rhea cohort team the and to a detailed analysis for by the Rhea study the Rhea website details of to of and has for All be by the Rhea study on a and we to to within 2 months. The Rhea cohort is a pregnancy cohort early life nutritional, environmental and social stressors may child health and development. Pregnant women (n = in the of Heraklion in Crete, Greece, were recruited at the time of the first major ultrasound (median 12 weeks of from February 2007 to 2008. of exposure and health outcomes was through from pregnancy until 6 years of age. assessment was conducted by trained and staff for the and third trimesters of and for the child 18 4 years, 6 years). The outcomes of are fetal and postnatal obesity, and behavioural asthma and The early life stressors are diet, environmental organochlorine phenols, heavy and water and extensive with repeated measurements has been created including urine, plasma, serum, extracted DNA and and and for data and collaborations are interested the and We all funding for the Rhea the European and The the for Research and in Greece and the Research Committee of the University of Crete, no of on analyses or We are to all the families who in the Rhea the research assistants and psychologists for and and the whole Rhea including research and for their and their in the of the study. We for the Rhea DNA We also for in the of Figure of