Cohort Profile: The Children’s Health in London and Luton (CHILL) cohort

The Children’s Health in London and Luton (CHILL) cohort was established to investigate the impact of London’s Ultra Low Emission Zone (ULEZ) on children’s health. Key strengths of CHILL include: the parallel prospective cohort, natural experimental design in which children living in London (exposed to the ULEZ) and children living in Luton (not exposed to a ULEZ or other major air-pollution control measure) are followed over time and compared; its large size (compared with similar studies) and ethnic diversity; high-resolution air pollution exposure data; and objective physiological measurements of lung function. In total, 3414 participants (aged 6–9 years; 52% female) were recruited in central London (intervention site, n = 1664) and Luton (comparator site, n = 1750). Baseline health assessments took place in 2018–19, prior to implementation of the ULEZ, with four annual follow-up assessments. Attrition ranged from 6.9% in the first follow-up year to 33.8% in the final year, with 1590 participants remaining after 5 years (n = 754 London, n = 836 Luton). Data were collected by using questionnaires, health assessments, and physical-activity monitoring. Study outcomes assessed include lung-function growth, physical activity, transport use, cognitive function and mental health indicators, healthcare utilization, personal and socioeconomic health determinants, and air-pollution exposures. The CHILL study welcomes collaborations and will consider all reasonable requests for data access. Requests, including a description of planned projects, should be sent to the corresponding author [email protected] at Queen Mary University of London, London, UK. Air pollution contributes to >8.1 million annual deaths worldwide and poses the largest environmental risk to health [1, 2]. This burden disproportionately affects the most disadvantaged and vulnerable members of society [3]. In Europe, nearly 800 000 excess deaths per year are attributed to poor air quality in densely populated urban areas [4] while, in the UK, most under-5s, young adults, and poorer households live in areas with the highest concentrations of traffic-related air pollution (TRAP) [5]. As their lungs are still developing, children are particularly sensitive to the harmful effects of air pollution [6] and concern has increased over recent years that long-term exposures can cause developmental delays with consequences to children’s health across the life course [7]. Exposure to TRAP has been associated with reduced lung function in primary-school children in London [8], Europe [9], and the USA [10]. Failure to achieve maximal lung-function growth by early adulthood is associated with increased risk of chronic respiratory disease and early mortality [11]. The imperative to reduce children’s exposure to air pollution [12], particularly in disadvantaged communities, is reflected in stringent clean air targets that were recently updated by the World Health Organization [13]. Clean air zones (CAZs), in which traffic control measures are employed to improve air quality, are an important tool in the effort to meet these targets and are being implemented in cities around the world [14]. The ultra low emission zone (ULEZ) in London is a CAZ that penalizes the entry of high-polluting vehicles (based on European class emission standards—see Supplementary Figure S1) and was introduced in 2019 with the aim of improving air quality and public health [15]. The Children’s Health in London and Luton (CHILL) cohort was established in 2018–19 to assess the impact of the ULEZ on children’s health and development by using a prospective parallel cohort design. The primary objective of the study is to investigate the impact of the ULEZ on lung-function growth. Secondary objectives will consider the impact of the ULEZ on air quality, respiratory health, physical activity, travel habits, health-related quality of life, and health costs not related to the National Health Service (NHS). Sub-studies undertaken alongside the CHILL study will investigate: the impact of air quality on cognitive development and mental health; the impact of air quality on epigenetics; exposure to heavy metals from TRAP; value for money of the ULEZ by undertaking a cost–consequence analysis. The CHILL study protocol has been published elsewhere [16]. An overview of the timeline and data collection is provided in Figure 1. The CHILL study was supported by a National Institute for Health Research (NIHR) Public Health Research grant (number 16/139/09) with additional funding for sub-studies by the NIHR Collaboration for Leadership in Applied Health Research (CLAHRC) North Thames, Barts Charity, The Mayor of London, and the NIHR Applied Research Collaboration (ARC) North Thames. Study overview. Health outcome data were collected annually, beginning in the 2018/19 academic year (prior to implementation of the ULEZ) and continuing for 4 follow-up years until 2022/3. The thick vertical lines indicate the implementation of the initial ULEZ area in central London (A) and the first expansion of the scheme to cover 380 km2 of inner London (B) (outlined on corresponding maps). GP, general practice (primary care). *Participants provided blood-spot samples in either the third or fourth follow-up health assessment. The cohort comprises children who are attending primary schools within or on the boundary of the initial ULEZ area in central London (see Figure 1) or in the Luton/Dunstable area (referred to as “Luton” for simplicity). Luton was selected as the comparator site due to its similarity to central London in terms of population demographics and mixture of local air-pollutant sources, while having no plans to introduce any form of CAZ and being sufficiently distant from London to minimize contamination from the ULEZ [17]. Ethical approval for the study was given by Queen Mary University of London Ethics of Research Committee (2018/08) and NHS West of Scotland Research Ethics Committee 4 (22/WS/0065). Written informed consent was obtained from the participants’ parent/carer. Participants were also asked for their verbal assent. All state primary schools in the study areas were approached, with 44 schools in central London and 40 in Luton (67% and 71% of those invited, respectively) agreeing to participate. We invited children in school Years 2, 3, and 4 (aged 6–9 years) to take part, except in large schools (more than three classes per year group), in which only children in Year 3 were invited. Study packs were delivered to parents in children’s school bags (including information sheets, a consent form, and a parental questionnaire). Sample-size requirements [16] necessitated a target of 3200 participants (1600 each in London and Luton) and 40 schools from each site. Of the children invited, signed consent forms were returned by 1664 in London (37%) and 1750 in Luton (36%). Children participating in the main study were also invited to take part in the CHILL sub-studies. Recruitment is summarized in Supplementary Figures S2 and S3. Children can perform spirometry from age 6 years [18], hence we recruited those aged 6–9 years at baseline to allow 2–3 years of follow-up within primary schools (to age 11 years). Initially, we did not intend to follow up participants after they had moved to secondary school; however, due to the disruption caused by the COVID-19 pandemic, we added a fourth year of follow-up for all participants, including those attending secondary schools. Four annual follow-up measurements have been taken since the baseline in 2018/19 (see Figure 1). Participation and attrition rates are provided in Table 1. Attrition was low in the first 2 years of follow-up (6.9% and 3.4%), but increased thereafter (21.8% and 33.8%), driven mainly by the movement of children from primary to secondary school. We were not able to trace some children once they had left primary school or visit all secondary schools to which participants had moved. Children who were lost to follow-up were on average slightly older than those who were retained throughout the study, reflecting the increased attrition of those moving to secondary school. Otherwise, their characteristics were similar (see Supplementary Table S1). Participation in the main CHILL study and sub-studies by year of data collection. Values are n (% of total recruited/retained), unless otherwise stated. FY, follow-up year; FEV1, forced expiratory volume in 1 second; FVC, forced vital capacity; GPS, global positioning system; cohort attrition is proportion of previous year’s participants who were lost to follow-up. Percentage of participants present at school for health assessment. Percentage of participants who attempted spirometry. Participation in the main CHILL study and sub-studies by year of data collection. Values are n (% of total recruited/retained), unless otherwise stated. FY, follow-up year; FEV1, forced expiratory volume in 1 second; FVC, forced vital capacity; GPS, global positioning system; cohort attrition is proportion of previous year’s participants who were lost to follow-up. Percentage of participants present at school for health assessment. Percentage of participants who attempted spirometry. During the second follow-up year (2020–1), our data collection was significantly shortened (to 4 months, April–July 2021) due to the COVID-19 pandemic. Annual parental questionnaire returns decreased over time, from 92.1% at baseline to 48.6% in the final follow-up year (Table 1). Returns were particularly low (38.7%) for the second follow-up year, in which questionnaires were mailed to participants due to the COVID-19 pandemic. Data were collected directly from children via health assessments, bio-samples, accelerometry, and questionnaires, and indirectly via questionnaires completed by their parents and teachers. Annual health assessments took place at the children’s schools and were performed by trained CHILL team members. Height, seated height, and weight were measured, and children completed a verbal questionnaire regarding asthma, use of inhalers, and their mode of transport to school. Lung function was assessed via pre- and post-bronchodilator spirometry, using Vitalograph 6000 Alpha Touch Spirometers (Vitalograph, Buckingham, UK) and following American Thoracic Society and European Respiratory Society guidelines [19]. Recorded measurements included forced expiratory volume in one second (FEV1), forced vital capacity (FVC), FEV1/FVC ratio, peak expiratory flow, and maximal mid-expiratory flow between 75% and 25% of the FVC. Parents were asked to complete an annual questionnaire coincidently with the health assessment (see Supplementary material). To assess physical activity, participants were given Actigraph triaxial accelerometers (Actigraph wGT3X-BT, Pensacola, Florida, USA) to wear for the week following the health assessment. A subset of participants were also given GPS monitors (QStarz BT-Q1000XT, Taipei, Taiwan). Physical-activity data were collected at baseline and during the first three annual follow-up health assessments (see Figure 1). Saliva samples were collected at the first and third follow-up assessments, using Oragene-DNA collection kits (OG-DNA). DNA methylation analysis will be performed by using the Infinium MethylationEPIC array to identify epigenetic signatures associated with short- and long-term air-pollution exposures. Finger-prick blood samples were collected at the third or fourth follow-up assessments and will be used to analyse heavy-metal exposures (lead, arsenic, and cadmium) via inductively coupled plasma mass spectrometry. Cognitive function and mental health symptoms were assessed at additional school visits between November 2019 and July 2022 (see Figure 1 and Supplementary material). Cognitive function measurements were obtained via interactive tasks performed on tablet computers (Lenovo ThinkPad Helix type 20CG or 20CH) by using a digital stylus or finger touch for input [20]. For the assessment of mental health symptoms, children completed the Revised Children’s Anxiety and Depression questionnaire (validated for self-completion for 8- to 18-year-olds) [21] and their class teachers completed the Strengths and Difficulties questionnaire. To assess the cost-effectiveness of the ULEZ in terms of healthcare utilization, primary-care records and hospital-episode statistics will be collected for participants with specific consent (requested at recruitment and confirmed in the final year of the study; see Supplementary material). Outcomes from the sub-studies will be published separately. To assess the impact of air quality on all study metrics, monthly and annual modeled air-pollutant exposures at the level of individual participant home addresses will be estimated for the criterion pollutants [i.e. those with legal standards—nitrogen dioxide (NO2), ozone (O3), particulate matter with an aerodynamic diameter of <10 microns (PM10) and <2.5 microns (PM2.5); see Supplementary material]. The characteristics of the London and Luton cohorts are shown in Table 2. Overall, slightly more girls were recruited than boys (52.1% vs. 47.9%), with the difference in proportions being more marked for London than for Luton. Differences were observed in the ethnicity of the two cohorts, with higher proportions of White and Asian/Asian British children recruited in Luton and higher proportions of Black/Black British and Mixed children recruited in London, reflecting the underlying populations (see Supplementary Tables S2A and S2B) [22, 23]. Height, weight, and body mass index (BMI) were similar between the two cohorts at baseline. The Index of Multiple Deprivation (IMD) was lower in London than in Luton (3.59 ± 1.70 vs. 4.03 ± 2.06), indicating higher deprivation in London (see Supplementary Figure S4). Age, sex, ethnicity, BMI, and IMD will be included as covariates in the final analysis. Characteristics of the cohort at baseline for all participants and by site. Data are presented as mean and standard deviation for continuous variables and as frequency and percentage for categorical variables. BMI, body mass index. Percentage of those who participated in the health assessment. Percentage of those who reported inhaler use. Percentage totals add to >100 because more than one answer could be given. Percentage of those who returned a completed parent questionnaire. Blue inhaler is a rescue inhaler (taken in response to symptoms), usually salbutamol or Ventolin. Brown inhaler is an inhaled corticosteroid (preventative), usually beclomethasone. Characteristics of the cohort at baseline for all participants and by site. Data are presented as mean and standard deviation for continuous variables and as frequency and percentage for categorical variables. BMI, body mass index. Percentage of those who participated in the health assessment. Percentage of those who reported inhaler use. Percentage totals add to >100 because more than one answer could be given. Percentage of those who returned a completed parent questionnaire. Blue inhaler is a rescue inhaler (taken in response to symptoms), usually salbutamol or Ventolin. Brown inhaler is an inhaled corticosteroid (preventative), usually beclomethasone. The mean annual air-pollution concentrations at the roadside and urban background sites within the ULEZ and in Luton for 2018 and 2019 are given in Table 3 (Supplementary Table S3 provides details of specific monitoring data used) [24–28]. Roadside NO2 concentrations (within 1–5 m of a busy road) were markedly higher within the ULEZ than in Luton, but this difference was strongly influenced by the kerbside sites (within 1 m of the kerb of a busy road) included in the London measurements, particularly Marylebone Road. Kerbside data were not available for Luton. Background NO2 data from an automated network were not available for Luton, but data from diffusion-tube measurements [28] indicated that background NO2 concentrations were also higher within the ULEZ than in Luton, albeit by a smaller margin than at the roadside. Measured annual mean pollutant concentrations from selected automatic air-quality-monitoring sites.a Data are presented as mean and SD. The N value is the number of monitoring sites included in mean. NO2, nitrogen dioxide; PM10, particulate matter with an aerodynamic diameter of <10 microns; PM2.5, particulate matter with an aerodynamic diameter of <10 microns. Urban background NO2 data for Luton are from diffusion-tube measurements; NA, data not available. Data sources: [26–30]. Measured annual mean pollutant concentrations from selected automatic air-quality-monitoring sites.a Data are presented as mean and SD. The N value is the number of monitoring sites included in mean. NO2, nitrogen dioxide; PM10, particulate matter with an aerodynamic diameter of <10 microns; PM2.5, particulate matter with an aerodynamic diameter of <10 microns. Urban background NO2 data for Luton are from diffusion-tube measurements; NA, data not available. Data sources: [26–30]. Higher roadside particulate concentrations were also observed within the ULEZ with Luton, these were background concentrations were similar at the two Background measurements were not available for Luton. data for NO2, PM2.5, and PM10, for sites are provided in the Supplementary (Supplementary Figure Children’s and inhaler use were similar between London and Luton at while exposure to was higher in Luton than in London vs. (Table respiratory symptoms (Table health-related quality of life (Supplementary Table and health-related costs (Supplementary Tables and were similar between London and Luton. Children’s mode of travel to school markedly between the two with more children to school via or or public transport in London with in Luton vs. vs. more children by or in Luton vs. of participants (n = attempted spirometry, with most of the primary outcome n = who did not were either or to spirometry. pre- and post-bronchodilator was for of participants (n = the of variables are given as and as of the value (see Table on height, weight, and ethnic as by using the Lung lung function were similar between London and Luton ± vs. ± ± vs. ± spirometry measurements at baseline for the cohort and by site. Data are presented as mean and standard FEV1, forced expiratory volume in 1 second; FVC, forced vital capacity; peak expiratory maximal mid-expiratory in Percentage of value on height, weight, and ethnic as by using the Lung pre- to Percentage in value spirometry measurements at baseline for the cohort and by site. Data are presented as mean and standard FEV1, forced expiratory volume in 1 second; FVC, forced vital capacity; peak expiratory maximal mid-expiratory in Percentage of value on height, weight, and ethnic as by using the Lung pre- to Percentage in value In the CHILL cohort comprises a large of children living in disadvantaged areas of London and Luton with of The cohort is for lung-function data and with parallel of air We have a parallel prospective cohort, natural design. To our this is in the of a air-pollution The strengths of this using a comparator include the to for in air quality from and to at for including the COVID-19 pandemic. The use of natural in the of population health a design is has been to be the only to to regarding the study outcomes from physiological parental and and across and of that health. The study from a high-resolution and to air-pollution exposures over the study In London, this is supported by monitoring of the criterion as as at roadside and background as summarized by the London Air This provides data to the that have in London following the of the The monitoring in Luton is but has throughout the of the The cohort provides a on which to add including to and additional outcomes to additional and the impact of of the A of the study is attrition of the cohort over To allow for a annual of participants was the used to baseline recruitment targets [16]. We an annual attrition of the total to follow-up being by the of data collection. A of our study is the between the and comparator the populations of central London and Luton have similar air pollution and monitoring are Luton has a similar of local air-pollutant but lower pollutant at roadside monitoring was in Luton at the beginning of the study with these Luton as a control site for our study, a to central London and to the effects of traffic pollution on health The CHILL study is to with other We will consider all reasonable requests for data access. should be with a description of the to the corresponding author and to [email protected] at Queen Mary University of London, London, UK. Ethics approval for on participants has been given by Queen Mary University of London Research Ethics Committee (2018/08) and NHS West of Scotland Research Ethics Committee 4 (22/WS/0065). All were in with guidelines and forms were completed by and children at Written informed consent was obtained from the participants’ parent/carer. Participants were also asked for their verbal assent. The study has been on We are to all participating and children for their and For the of the author has a to any The study was by and The first of the was by and with the on air quality being by provided by All on previous of the All and the final will as for the Supplementary data is available at of This was supported by National Institute of Health Research (NIHR) Public Health Research number 16/139/09) with additional funding by NIHR North Thames, Barts Charity, and NIHR North Thames. is by the Research from the NIHR Barts Research is supported by the NIHR North Thames. and were supported by the NIHR Health Research in and and and and a between the Health and is also by Research and Research The in this are those of the and not those of the NIHR or the of Health and no in the design of the study, the and of the in the The used during the study are available from the corresponding author reasonable