1 BACKGROUND

The early life programming concept was originally proposed to explain the link between maternal nutritional deficits during pregnancy, low birth weight and associated risks for obesity, diabetes and cardiovascular disease in lifespan.1 These initial observations were the basis of the developmental origins of health and disease (DOHaD) paradigm.2, 3 Population-based prospective pregnancy and birth cohorts provide one of the most appropriate study designs to investigate early life hazards at critical time points of prenatal and postnatal human development.

According to the Global Asthma Report 2018, asthma affects over 330 million people worldwide and its prevalence is rising.4 Moreover, asthma is the most common chronic disease among children.5 In addition, approximately 30%–40% of the world's population is affected by one or more allergic conditions, with vast personal, social, and economic costs.6 Despite some advances in treatment, this condition continues to be a public health concern, still killing as many people as malaria.7 Primary prevention is thus essential to reduce the burden of asthma.

The Nutrition in Early Life and Asthma (NELA) study was launched to try and offer information which may produce primary prevention interventions. The main objective of the NELA study was to investigate whether maternal obesity/adiposity and foetal growth; prenatal and postnatal nutrition; outdoor air pollution; endocrine disruptors; and maternal psychological stress contribute causally to the development of asthma.

2 METHODS 2.1 Cohort design and population

The NELA study is a prospective, population-based, maternal-child, birth cohort study which was set up in 2015 at the “Virgen de la Arrixaca” University Clinical Hospital of Murcia, a south-eastern Mediterranean region of Spain. Randomly selected pregnant women who fulfilled the inclusion criteria were invited to participate at the time of the follow-up visit at the aforementioned hospital for routine foetal anatomy scan at 19–22 gestation week, at the Maternal-Fetal Medicine Unit of the hospital from March 2015 to April 2018. During that period, two mornings per week, one of the two principal investigators invited to participate as many mothers as possible from those attending the unit and meeting the inclusion criteria.

Inclusion criteria were: Caucasian and Spanish origin; 18–45 years of age; living in Health Area I (suburban and rural) or in certain districts of Health Areas VI and VII (mainly urban) of the Region of Murcia (Figure 1); planning to live in the area of study during at least 2 years; singleton pregnancy; spontaneous conception; intention to deliver in the aforementioned hospital; and normal ultrasound findings at the time of the visit (no major foetal malformations). Exclusion criteria included: chronic disease in the mother, such as pregestational diabetes mellitus or other major endocrine disorders, pregestational hypertension, autoimmune disease, or cancer; and verbal communication problems.

Sampling areas of the province of Murcia. The roman numbers indicate heath areas

2.2 Data collection and follow-up time points

Mothers had three follow-up visits, one in the second trimester of pregnancy (between 20 and 24 gestation week), one in the third trimester (between 32 and 36 gestation week), and one at delivery. Children have been followed up at birth, and at 3 and 18 months of age so far. They are being followed up at 5 years and will be followed up further on (Figure 2). Table 1 (mothers) and Table 2 (children) show the information obtained or planned at the different visits from mothers and children respectively.

Summary of data collection flow in the NELA birth cohort study. Due to SARSCov-2 pandemic, only information by telephone is being collected at the follow-up visit of five years

TABLE 1. Summary of the information obtained by questionnaires and measurements from mothers at the different follow-up visits Pregnancy Delivery Children's age 20–24 wk. 32 wk 3 mo. 18 mo. 5 yr. 7 yr. Information Obstetric & medical history ▲ Asthma/allergy ▲ Infection disease ▲ ▲ ▲ Anxiety/depression ▲ ▲ ▲ Tobacco smoking ▲ ▲ ▲ ▲ ▲ ▲ Alcohol/drugs consumption ▲ ▲ ▲ Diet & supplements ▲ ▲ Physical exercise ▲ ▲ Medicine intake ▲ ▲ Occupational hazards ▲ Endocrine disruptors ▲ ▲ Outdoor air pollution ▲ ▲ ▲ Measurements Weight ▲ ▲ ▲ Height ▲ Skin folds ▲ ▲ Body composition ▲ ▲ Blood pressure ▲ ▲ Lung function ▲ Skin prick test ▲ eVOCsa ▲ a Exhaled volatile organic compounds. TABLE 2. Summary of the information obtained by questionnaires and measurements from children at the different follow-up visits Time point Birth 3 mo. 18 mo. 5 yr. 7 yr. Information Respiratory problems ▲ ▲ ▲ ▲ Atopic eczema ▲ ▲ ▲ ▲ Asthma/allergy ▲ ▲ ▲ ▲ Skin prick test ▲ Cognitive development ▲ ▲ Breast feeding ▲ ▲ Indoor tobacco smoking ▲ ▲ ▲ ▲ Diet & supplements ▲ ▲ ▲ ▲ Pet contact ▲ ▲ ▲ ▲ Medicine intake ▲ ▲ ▲ ▲ Vaccinations ▲ ▲ ▲ ▲ Physical exercise ▲ ▲ ▲ Measurements Weight ▲ ▲ ▲ ▲ ▲ Height ▲ ▲ ▲ ▲ ▲ Head circumference ▲ ▲ ▲ ▲ Waist circumference ▲ ▲ ▲ ▲ Skin folds ▲ ▲ ▲ ▲ Blood pressure ▲ Th1/Th2a & cytokines ▲ Lung function ▲ ▲ Outdoor air pollution ▲ ▲ ▲ ▲ ▲ eVOCsb ▲ a T-helper (Th). b Exhaled volatile organic compounds. 2.3 Specific objectives - To investigate whether maternal adiposity and whether foetal growth patterns at 12, 20 and 32 gestation week are associated with lung function and volatile organic compounds in exhaled air (eVOCs) at 3 months of age; and with subsequent asthma. - To establish whether factors acting in the mother during pregnancy are associated with changes in lung function and/or in pattern(s) of eVOCs at 3 months; and to subsequent asthma. - To establish whether any of the potential associations is mediated by a change in the Th1/Th2 immunological profile. - To investigate disturbances in the maternal and offspring microbiome in relation to developmental influences and their role in subsequent occurrence of asthma. - To try and detect any epigenetic changes in offspring at birth which could mediate the potential effects of the aforementioned factors acting in the mother during pregnancy. - To build up a unique biobank sample collection with maternal and child biological samples. 2.4 Main outcome variables

The main outcome variables are lung function at three months of age of the children; and asthma (symptoms and/or diagnosis) and related phenotypes at any point of the follow-up visits (Figure 3).

Diagram of the different research areas covered by the nutrition in early life and asthma (NELA) cohort and their interactions

2.4.1 Lung function

In those children whose parents authorised the needed sedation for the technique (chloral hydrate 80–100 mg/Kg), lung function tests were performed at the Paediatric Respiratory unit of the hospital. Forced vital capacity (FVC), forced expiratory volume at 0.5 seconds (FEV0.5) and forced expiratory flows at 75 and 25%–75% of FVC (FEF75 and FEF25-75) were obtained from maximal expiratory volume curves by means of the raised volume rapid thoracic compression (RVRTC) technique according to the ERS/ATS consensus statement.8 Oxygen saturation was continuously measured until recovery from sedation. Tests were performed with a Master-Screen BabyBody plethysmograph (Jaeger, Germany) and raised volume was achieved by coupling a Neopuff infant resuscitator (Fisher & Paykel Healthcare, New Zealand) to the facemask. According to published regression formulas corrected for the Jaëger equipment, z-scores of FVC, FEV0.5, FEF75 and FEF25-75 were calculated.9 Although RVRTC needs sedation and is time-consuming, we preferred it to the multiple breath washout with lung clearance index (MBW-LCI), which can be performed in non-sedated infants because RVRTC measures lung flows and volumes very similarly to spirometry, while MBW-LCI measures ventilation inhomogeneity. This and other lung function techniques, although providing measurements of lung function, are not directly comparable to spirometry.

2.4.2 Symptoms and diagnoses

Symptoms of potential asthma, allergic rhinoconjunctivitis and atopic eczema and their medical diagnoses by the mothers’ or children’ doctors were surveyed in all follow-up visits. Symptoms were asked for using questions from the “Estudio Internacional de Sibilancias en Lactantes” (EISL)10 or Global Asthma Network (GAN)11 questionnaires, depending on the age, to mothers and children. Questions on wheezing in the different visits will allow to categorise wheezing children into the classical phenotypes of transient early, late-onset and persistent to know whether risk or protective factors vary across phenotypes.

2.5 Developmental influences 2.5.1 Maternal adiposity and foetal growth

Maternal abdominal ultrasounds and Doppler examinations (Voluson 730 Expert, GE Medical Systems, Austria) were performed, recorded and used to obtain foetal biometry at 20 and 32 gestation weeks. Foetal weight was estimated from those measurements.12 Amniotic fluid and placental location and structure were also assessed. Doppler pulsatility index of umbilical and maternal uterine arteries was also measured. Abdominal ultrasound was used to measure subcutaneous fat thickness.13 Ultrasound scan at 12 gestation week was obtained from clinical records and gestational age was calculated from the foetal crown-rump length.14

2.5.2 Maternal body composition

Maternal body composition throughout pregnancy was measured by bioelectrical impedance (Bodystat Quadscan 4000, Isle of Man, UK). Somatotype was obtained from 10 anthropometric variables: weight, height, skin folds (triceps, subscapularis, supraspinatus and calf) and diameters (arm and calf). Medidep (Vigo, Spain) was used to calculate the percentages of bone, lean and fat mass, together with the components of endomorphy, mesomorphy and ectomorphy.15 Somatotype will be represented as somatochart.

2.5.3 Maternal diet during pregnancy

Diet information was collected at 20 gestation week, using a Food Frequency Questionnaire (FFQ) based on a validated questionnaire16 administered by trained interviewers. The FFQ has semi-quantitative (112) and qualitative (11) items. For each food item, the questionnaire asks how often, of nine possible intake frequency categories, the participants have consumed a particular amount of food from the beginning of pregnancy until the time of the interview; and standard units or reference serving sizes are specified. The intake frequency for each food item was converted to the average daily intake. Then, nutrient values and energy intakes were obtained from different sources.17, 18

To evaluate the degree of adherence to a Mediterranean Diet during pregnancy, two scores are being used: Alternative Mediterranean Diet score19 (aMED) and Relative Mediterranean Diet score20 (rMED) both a modified version of the Mediterranean Diet Score (MDS).21 Other diet patterns will be also explored.

2.5.4 Nutritional biomarkers

Maternal and offspring vitamin levels were quantified in maternal serum at 24 gestation week and in cord blood serum; 25-hydroxyvitamin D [25(OH)D] was measured by direct competitive immunoluminometric assay using coated magnetic microparticles in a LIAISON® XL automated analyser (DiaSorin S.p.A.). Both vitamin A and Vitamin E were extracted (CI&C GmbH) and quantified by HPLC (Waters, Barcelona, Spain). Fatty acids were analysed in maternal plasma at 24 gestation week in both arterial and venous cord blood to detect disturbances in their placental transfer and foetal fat uptake. They were quantified by gas chromatography on a Hewlett-Packard 6890 as previously described22 (Agilent Technologies, Inc) equipped with a SP-2560 capillary column (60 m × 0.25 mm id × 0.15 μm; Supelco, SIGMA-Aldrich). Polyamines putrescine, spermidine and spermine were quantified by HPLC methods in both maternal serum at 24 gestation week and venous cord blood serum. Polyamines were derivatised to dansyl compounds and quantified using a reverse-phase column (Nova-Pak C18). A two-phase gradient, and detection by immunofluorescence was used.23

2.5.5 Residential outdoor air pollution

To overcome the limitation of having too few air quality stations,24 we followed the air quality Directive 2008/50/EC criteria for using high-resolution air quality models for complementing on-site observations. Thus, the Weather Research and Forecasting (WRF)+CHIMERE modelling system was developed for generating the outdoor concentrations for air pollutants during the entire duration of the NELA cohort study. To include the pregnancy period, simulations cover from nine months prior to the first enrolled women (July 2014) to the end of 2018.

The regional meteorological model Advanced Research Weather Research and Forecasting (WRF-ARW) Model v3.9.1.25 was used to provide the meteorology to the chemistry transport model. WRF is driven every six hours by ERA-Interim reanalysis and has been coupled off-line on an hourly basis to CHIMERE chemistry transport model.26 WRF fields are interpolated to CHIMERE working grids. MELCHIOR2 gas-phase mechanism is implemented within CHIMERE. The chemistry transport model includes gas-phase chemistry and aerosols; and heterogeneous chemistry distinguishes among different chemical aerosol components. The physico-chemical options for the regional modelling system are summarised elsewhere.27

With respect to anthropogenic and natural emissions, a specific emission inventory was developed for the purpose of the NELA study. Biogenic emissions are coupled to WRF outputs. The model estimates hourly isoprene, monoterpene and other biogenic VOC emissions.

The final working resolution over the three target Health Areas of NELA recruitment is 0.5 km. The exposure to air pollution is derived at short term (15 days before birth) and long term (entire pregnancy and trimester-specific). Information on residential address was collected in the follow-up visits. A geo-codification and geo-referencing process were implemented so the exposure to air pollution was extracted for the corresponding time-period and the exact geo-location of the mother-sib pair. This will be further applied at other time points of follow-up.

Exposure to indoor air pollution is being assessed through questionnaires in the different visits.

2.5.6 Endocrine disruptors

Urine samples from the mothers were collected in the first visit. Samples were immediately frozen at −80°C. The main six organophosphate pesticide metabolites (dialkylphosphates) such as dimethylphosphate (DMP), dimethylthiophosphate (DMTP), dimethyl-dithiophosphate (DMDTP), diethylphosphate (DEP), diethylthiophosphate (DETP), and diethyldithiophosphate (DEDTP) will be analysed in due course of time using gas chromatography coupled with tandem mass spectrometry (GC -MS/MS). Additionally, as markers of prenatal exposure to endocrine disruptors, anogenital distance (AGD) and penile width were measured at the birth visit, as previously described.28

2.5.7 Psychological stress in the mother and child neurodevelopment

Psychological well-being during pregnancy was assessed by means of the Family Apgar test at 20 gestation week.29 Additionally, the Edinburgh postnatal depression scale,30 used to assess prenatal depression,31 was administered at 20 and at 32 gestation week. Furthermore, the State-Trait Anxiety Inventory (STAI) test was included in the 32 gestation week visit.32 Children were evaluated at 18 months of age by means of the Spanish version of the Bayley III scales for infant and toddler development.33 During the same visit, externalising and internalising behaviour of parents was assessed using the Spanish version of the Child Behavior Checklist (CBCL) scale.34 Their adaptive and social behaviour were also measured through the Behavior Assessment System-Third Edition (ABAS-III) scale.35

2.6 Mechanisms: Systems biology 2.6.1 Immune system phenotyping in the offspring

Measurements of immune system cells were performed by flow cytometry. Lymphoid cells were gated according to size (Forward Scatter, FSC) and granularity (Side Scatter, SSC) parameters. Frequency and absolute numbers of immune subpopulations Th1, Th2, Th17, Treg and ILC2 were obtained according to their cell surface markers.36, 37 Expression level of each marker was also analysed by using the mean fluorescence intensity data. Additionally, mononuclear cells from cord blood were isolated by density gradient centrifugation with Ficoll-Hypaque Plus (Amersham Biosciences).38

To obtain cord blood cytokine profiles, whole blood samples were incubated in the presence of specific immune stimulants or medium alone39:Concanavalin A, D. pteronyssinus and Olea europaea extracts, oligonucleotides containing CpG motifs, polyinosinic-polycytidylic acid (pl:C), peptidoglycan (PEG), lypopolisaccharide (LPS), and phytohaemaglutinin. Stimulants were chosen according to their capacity to induce innate and/or Th1, Th2 or Th17 cytokines in adults.40 After incubation, supernatants were collected and analysed for the presence of the following cytokines: IFN-alpha, IFN-gamma, IL-1beta, IL-2, IL-4, IL-5, IL-6, IL-10, IL-13, IL-17F, IL-23 and TNF-alpha. Concentrations of all these cytokines were measured simultaneously by the Luminex cytokine kit (Thermo Fisher Scientific) and run on a MAG-PIX (Luminex Corporation), equipped with xPONENT software (Luminex Corporation). Cytokine results were analysed by ProcartaPlex Analyst 1.0 Software (Thermo Fisher Scientific).

2.6.2 Breathomics

Breath sampling, exhaled breath analysis and data preprocessing have been detailed recently.41 Exhaled breath samples were collected from mothers and infants at three months of age in plastic free bags (QuintronTM and TedlarTM gas bags for children, and TedlarTM gas bags for mothers). Room air samples were also collected by an Easy-VOC syringe (Markes Int.TM). The air contained in the sample bags was transferred to adsorption tubes for trapping VOCs (Tenax TA, Markes Int.TM). The tube was sealed with brass caps fitted with PTFE ferrules and stored at 4 °C until analysis by thermal desorption coupled to chromatography-mass spectrometry (TD-GC/MS). Raw output data from GC/MS was preprocessed using a specifically created open code workflow in R language with functions from xcms, cliqueMS and eRah.41 VOCs were identified through National Institute of Standard and Technology (NIST) spectral library matching by means of their retention time and their respective mass spectrum. In addition, chromatographic standards were used for retention indexes computation. Special attention will be paid to compounds previously related to asthma.