Epidemiology and Biostatistics

Imperial College e-Science Networked Infrastructure (ICENI) is an end-to-end Grid middleware developed to allow transparent usage of Grid. It consists of both a service-oriented Grid middleware and an application toolkit, using a component-programming model to represent Grid applications. We utilise this infrastructure in the the e-Protein project, a Biotechnology and Biological Sciences Research Council pilot project, which provides structure-based annotation of proteins in the major genomes by linking remote resources at three different sites. The annotation utilises homology and fold recognition methods to assign protein structures to the proteomes. We have taken the approach of modularising applications into inter-connected components and running each application through a workflow manager, which executes the components in the specified manner. A web interface provides a user-friendly front end to the Grid system, where users can request their choice for the annotation and check previous annotations from the database.

The Grid holds great potential for users, software developers and resource owners. Users of the Grid are abstracted away from its complexity, while software developers are being provided with a rich middleware in which to develop their applications without the need for a large investment in resources. Resource owners are able to expose their resources, potentially for financial reward, without the need to invest in software. However, as most of the applications which currently exist pre-date the concept of the Grid, they lack the appropriate functionality and/or modularization to exploit it. In this paper we propose the use of a standards based job submission and monitoring system which is capable of deploying legacy applications onto existing resources within the Grid, environment containers for mapping applications into Grid applications, along with the use of web-based portals to expose these applications to the end user. We further propose that applications which comprise a number of legacy tasks can be handled through the use of a workflow enactment service submitting each of these tasks through the standards based job submission system. We exemplify our approach through the e-Protein project by taking their existing proteome annotation software and exposing it as a Grid service. We show how the use of this technique over the Grid can significantly reduce the makespan required for the application.

This paper summarizes the epidemiological evidence for adverse health effects associated with disinfection by-products (DBPs) in drinking water and describes the potential mechanism of action.

There appears to be good epidemiological evidence for a relationship between exposure to DBPs, as measured by trihalomethanes (THMs), in drinking water and bladder cancer, but the evidence for other cancers including colorectal cancer is inconclusive and inconsistent. There appears to be some evidence for an association between exposure to DBPs, specifically THMs, and little for gestational age/intrauterine growth retardation and, to a lesser extent, pre-term delivery, but evidence for relationships with other outcomes such as low birth weight, stillbirth, congenital anomalies and semen quality is inconclusive and inconsistent. Major limitations in exposure assessment, small sample sizes and potential biases may account for the inconclusive and inconsistent results in epidemiological studies. Moreover, most studies have focused on total THMs as the exposure metric, whereas other DBPs appear to be more toxic than the THMs, albeit generally occurring at lower levels in the water.

The mechanisms through which DBPs may cause adverse health effects including cancer and adverse reproductive effects have not been well investigated. Several mechanisms have been suggested, including genotoxicity, oxidative stress, disruption of folate metabolism, disruption of the synthesis and/or secretion of placental syncytiotrophoblast-derived chorionic gonadotropin and lowering of testosterone levels, but further work is required in this area.

Disinfection by-product (DBP) exposure during pregnancy may be related to reduced fetal growth, but the evidence is inconclusive and improved DBP exposure assessment is required. The authors conducted a nested exposure study on a subset (n ¼ 39) of pregnant women in the Born in Bradford cohort to assess validity of TCAA exposure assessment based on tap water sampling and self-reported water-use; wateruse questionnaire validity; and use of a one-time urinary TCAA biomarker. TCAA levels in urine and home tap water supply were quantified, and water use was measured via a questionnaire and 7-day diary, at 28 weeks gestation. Diary and urine measures were repeated later in pregnancy (n ¼14). TCAA level in home tap water supply was not correlated with urinary TCAA (0.18, P¼ 0.29). Cold unfiltered tap water intake at home measured by questionnaire was correlated with urinary TCAA (0.44, P¼ 0.007), but correlation was stronger still for cold unfiltered tap water intake reported over the 3 days prior to urine sampling (0.60, P o0.001). For unemployed women TCAA ingestion at home, derived from tap water sampling and selfreported water-use, correlated strongly with urinary TCAA (0.78, P o0.001), but for employed women the correlation was weak (0.31, P ¼0.20). Results suggest individual tap water intake is most influential in determining TCAA exposure variability in this cohort, and that TCAA ingestion at home is a valid proxy for TCAA exposure for unemployed women but less satisfactory for employed women.

There appears to be very good epidemiological evidence for a relationship between chlorination by-products, as measured by trihalomethanes (THMs), in drinking water and bladder cancer, but the evidence for other cancers, including colorectal cancer appears to be inconclusive and inconsistent. There appears to be some evidence for a relationship between chlorination by-products, as measured by THMs, and small for gestational age (SGA)/intrauterine growth retardation (IUGR) and preterm delivery, but evidence for other outcomes such as low birth weight (LBW), stillbirth, congenital anomalies and semen quality appears to be inconclusive and inconsistent.

Disinfection of drinking water is vital to protect the public against disease. However disinfectants such as chlorine react with organic matter in drinking water to produce a wide range of chemical disinfection by-products (DBPs) of potential health concern including haloacetic acids (HAAs). This thesis is an epidemiologic analysis investigating the relationship between prenatal exposure to HAAs in drinking water and adverse birth outcomes in "Born in Bradford", a large multi-ethnic prospective birth cohort study based in Bradford, England. It focuses on the understudied and as yet unregulated HAAs which are the second most prevalent class of chlorination DBPs in UK drinking waters. To assess exposure, area-level concentrations to three select HAAs (measured in drinking water samples newly collected for this study, modelled in time and space, and weighted to each cohort woman's specific trimester of pregnancy by postcode of residence) were combined with individual water consumption information collected via questionnaire at recruitment to the cohort. Despite the benefits of state-of-the-art exposure metrics and a large sample size, this study does not find any significant patterns of association between prenatal exposure to HAAs and either birth weight, being born term low birth weight or small-for-gestational age. Water consumption over the course of late pregnancy was further studied in a subset of cohort women. A small but significant increase in water consumption was reported, bearing in mind that both behaviour change over the third trimester of pregnancy and measurement error likely contributed to this effect. This research addresses some of the limitations of previous DBP studies in terms of exposure assessment and birth outcome definitions, and uniquely evaluates the variability of individual water consumption over time. It also identifies areas for future research and examines the importance of HAAs and birth weight-based outcomes in the larger research context. I hereby declare that the contents of this PhD thesis are all my own work and that where any material could be construed as the work of others it has been fully cited and referenced, and/or the appropriate acknowledgement given. May 2010 and November 2010, I collected the HiWATE water samples in Bradford. This involved coordinating sampling dates with Yorkshire Water, sending instructions and clean labelled vials containing preservatives by courier, and picking up the samples after collection. I was then charged with driving those samples from Bradford to Cranfield University for laboratory analyses. This work was made possible by the collaboration and expertise of various individuals. Dr Kees de Hoogh completed the automated geocoding of Born in Bradford participants' residence and work postcodes and addresses; I did the manual geocoding. Dr Hannah Slater advised me on linking women's addresses to water supply zones via ArcGIS (Chapter 4). Professor Nicky Best and Dr Juan Gonzalez Maffe helped me develop the WinBUGS models both for the HAA modelling (Chapter 4) and the repeat questionnaire study (Chapter 8), and advised me on the statistics relating to the main epidemiologic analysis (Chapter 7). Drs James Bennett and Léa Fortunato helped me in the initial stages of HAA modelling as well. The idea for the repeat questionnaire study came from Professor Mark Nieuwenhuijsen and Dr Mireille Toledano. I got ethics approval for the study in September 2010 and conducted the study between September and December 2010. No personal data were to leave the BiB office for confidentiality reasons. I therefore travelled to Bradford to access women's information, in order to prepare the mailings. I then collected the returned questionnaires, entered all data manually, and developed the statistical analysis plan. I kept in close contact with Dr Emily Petherick from the Born in Bradford team regarding the data extracts, and data cleaning issues.

ObjectivesTo construct UK ethnicity birth weight centiles (UK-EBWC) for gestational age and cut-offs for small for gestational age (SGA) for England and Wales and to evaluate the SGA misclassification using the UK centiles.DesignAnalysis of national birth data.ParticipantsAll live singleton births in England and Wales in 2006–2012, as recorded by the Office for National Statistics and birth registrations, linked with National Health Service into numbers for babies.Main outcome measuresBoth sex-specific and ethnicity-sex-specific birth weight centiles for gestational age, and ethnicity-sex-specific SGA cut-offs. Centiles were computed using the generalised additive model for location, scale and shape.ResultsOur sex-specific centiles performed well and showed an agreement between the expected and observed number of births below the centiles. The ethnicity-sex-specific centiles for Black and Asian presented lower values compared with the White centiles. Comparisons of sex-specific and ...

Bulky DNA adducts reflect genotoxic exposures, have been associated with lower birthweight and may predict cancer risk. We selected factors known or hypothesized to affect in utero adduct formation and repair and examined their associations with adduct levels in neonates. Pregnant women from Greece, Spain, England, Denmark and Norway were recruited in 2006-2010. Cord blood bulky DNA adduct levels were measured by the (32)P-postlabelling technique (n=511). Diet and maternal characteristics were assessed via questionnaires. Modeled exposures to air pollutants and drinking-water disinfection by-products, mainly trihalomethanes (THMs) were available for a large proportion of the study population. Greek and Spanish neonates had higher adduct levels than the Northern European neonates (median: 12.1 (n=179) vs. 6.8 (n=332) adducts per 10(8) nucleotides, p<0.001). Being from the Southern European countries, higher maternal body mass-index, delivery by Caesarean section, being boy, low ma...

Background: Studies of disinfection by-products in drinking water and measures of adverse fetal growth have often been limited by exposure assessment lacking data on individual water use, and therefore failing to reflect individual variation in DBP exposure. Methods: Pregnant women recruited to the Born in Bradford cohort study completed a questionnaire which covers water exposure. Information was collected on water consumption, showering, bathing and swimming. Water exposure data from a subset of 39 women of the cohort are described here. Results: Mean total tap water intake was 1.8 l/day, and women on average spent 146 minutes per week showering and bathing. Most tap water intake occurred at home (100% for unemployed, 71.8% for employed). Differences between age groups were observed for total tap water intake overall (p = 0.02) and at home (p = 0.01), and for bottled water intake (p = 0.05). There were differences between ethnic groups for tap water intake at home (p = 0.02) and total tap water intake at work (p = 0.02). Total tap water intake at work differed by income category (p = 0.001). Duration per shower was inversely correlated with age (Spearman's correlation -0.39, p = 0.02), and differed according to employment status (p = 0.04), ethnicity (p = 0.02) and income (p = 0.02). This study provides estimates of water exposure in pregnant women in a multi-ethnic population in the north of England and suggests differences related to age, employment, income and ethnicity. The findings are valuable to inform exposure assessment in studies assessing the relationship between DBPs and adverse birth outcomes.

Large-scale prospective cohort studies are invaluable in epidemiology, but they are increasingly difficult and costly to establish and follow-up. More efficient methods for recruitment, data collection and follow-up are essential if such studies are to remain feasible with limited public and research funds. Here, we discuss how these challenges were addressed in the UK COSMOS cohort study where fixed budget and limited time frame necessitated new approaches to consent and recruitment between 2009-2012. Webbased e-consent and data collection should be considered in large scale observational studies, as they offer a streamlined experience which benefits both participants and researchers and save costs. Commercial providers of register and marketing data, smartphones, apps, email, social media, and the internet offer innovative possibilities for identifying, recruiting and following up cohorts. Using examples from UK COSMOS, this article sets out the dos and don'ts for today's cohort studies and provides a guide on how best to take advantage of new technologies and innovative methods to simplify logistics and minimise costs. Thus a more streamlined experience to the benefit of both research participants and researchers becomes achievable.

Evidence for a relationship between trihalomethane (THM) or haloacetic acid (HAA) exposure and adverse fetal growth is inconsistent. Disinfection by-products exist as complex mixtures in water supplies, but THMs and HAAs have typically been examined separately. To investigate joint exposure at individual level to THMs and HAAs in relation to birth weight in the multi-ethnic Born in Bradford birth cohort. Pregnant women reported their water consumption and activities via questionnaire. These data were combined with area-level THM and HAA concentrations to estimate integrated uptake of THMs into blood and HAA ingestion, accounting for boiling/filtering. We examined the relationship between THM and HAA exposures and birth weight of up to 7,438 singleton term babies using multiple linear regression, stratified by ethnicity. Among Pakistani-origin infants, mean birth weight was significantly lower in association with the highest versus lowest tertiles of integrated THM uptake (e.g. -53.7...

The United Kingdom COhort Study of MObile phone uSe and health (UK COSMOS) is a prospective cohort study established to investigate the possible health effects associated with long-term use of mobile phones and other wireless technologies, to inform public health policy in the UK and beyond. UK COSMOS is part of the international COSMOS cohort study consortium (the UK, Sweden, The Netherlands, Finland, Denmark and France) on mobile phones and health, which has over 300 000 study participants across the six partner countries. 1 Details of consortium partners are given at [www.ukcosmos.org]. A large prospective cohort study of mobile phone users with long-term follow-up has been recommended as a high priority by the World Health Organization. This reflects both scientific uncertainty and public concern regarding possible health effects of mobile phone use. Radiofrequency electromagnetic fields have recently been classified as possibly carcinogenic to humans (Group 2B) on the basis of limited evidence concerning risk of brain cancers. Results from the Interphone study, the largest case-control study on brain cancer to date, suggested possible increased risks of glioma at the highest levels of mobile phone use, but interpretation was unclear and the possible health effects of long-term heavy use of mobile phones remain uncertain. The COSMOS study was designed to investigate this question while addressing limitations of previous studies. Its cohort study design, with prospective exposure assessment, is less prone to potential selection and recall biases associated with case-control studies such as Interphone. Unlike the case-control approach, it addresses a wide range of disease outcomes of importance for public health in one investigation, including neurodegenerative disease, stroke and depression which have rarely been studied; previous studies have focused on few outcomes, mainly tumours of the brain and head. Large long-term cohort studies have huge potential to increase understanding of environmental exposures and disease and thus improve the health of current and future generations, but require sustained investment from funders and continued involvement by participants. To maximize the value of this investment and the benefit to public health, UK COSMOS is not only collecting data to address the question of mobile phones and health, but also has obtained extensive data on wider environmental exposures, lifestyle and demographics, so as to address a wide range of environment and health questions. With over 100 000 participants, UK COSMOS is the UK's fourth largest cohort study 6 and the largest cohort in the international COSMOS consortium. The UK COSMOS study protocol and subsequent amendments

In the "Results" section of the abstract, "median, 12.1 (n = 179) vs. 6.8 (n = 332) adducts per 108 nucleotides, p < 0.001" should have read "median, 12.1 (n = 179) vs. 6.8 (n = 332) adducts per 10 8 nucleotides, p < 0.001." The published article has been updated online. EHP regrets this error.

Objective To investigate the relation between exposure to both air and noise pollution from road traffic and birth weight outcomes.Design Retrospective population based cohort study.Setting Greater London and surrounding counties up to the M25 motorway (2317 km2), UK, from 2006 to 2010.Participants 540 365 singleton term live births.Main outcome measures Term low birth weight (LBW), small for gestational age (SGA) at term, and term birth weight.Results Average air pollutant exposures across pregnancy were 41 μg/m3 nitrogen dioxide (NO2), 73 μg/m3 nitrogen oxides (NOx), 14 μg/m3 particulate matter with aerodynamic diameter &lt;2.5 μm (PM2.5), 23 μg/m3 particulate matter with aerodynamic diameter &lt;10 μm (PM10), and 32 μg/m3 ozone (O3). Average daytime (LAeq,16hr) and night-time (Lnight) road traffic A-weighted noise levels were 58 dB and 53 dB respectively. Interquartile range increases in NO2, NOx, PM2.5, PM10, and source specific PM2.5 from traffic exhaust (PM2.5 traffic exhaust)...

This paper summarizes the epidemiological evidence for adverse health effects associated with disinfection by-products (DBPs) in drinking water and describes the potential mechanism of action. There appears to be good epidemiological evidence for a relationship between exposure to DBPs, as measured by trihalomethanes (THMs), in drinking water and bladder cancer, but the evidence for other cancers including colorectal cancer is inconclusive and inconsistent. There appears to be some evidence for an association between exposure to DBPs, specifically THMs, and little for gestational age/intrauterine growth retardation and, to a lesser extent, pre-term delivery, but evidence for relationships with other outcomes such as low birth weight, stillbirth, congenital anomalies and semen quality is inconclusive and inconsistent. Major limitations in exposure assessment, small sample sizes and potential biases may account for the inconclusive and inconsistent results in epidemiological studies. Moreover, most studies have focused on total THMs as the exposure metric, whereas other DBPs appear to be more toxic than the THMs, albeit generally occurring at lower levels in the water. The mechanisms through which DBPs may cause adverse health effects including cancer and adverse reproductive effects have not been well investigated. Several mechanisms have been suggested, including genotoxicity, oxidative stress, disruption of folate metabolism, disruption of the synthesis and/or secretion of placental syncytiotrophoblast-derived chorionic gonadotropin and lowering of testosterone levels, but further work is required in this area.

To investigate the relation between exposure to both air and noise pollution from road traffic and birth weight outcomes. Retrospective population based cohort study. Greater London and surrounding counties up to the M25 motorway (2317 km 2 ), UK, from 2006 to 2010. PARTICIPANTS 540 365 singleton term live births. Term low birth weight (LBW), small for gestational age (SGA) at term, and term birth weight. Average air pollutant exposures across pregnancy were 41 μg/m 3 nitrogen dioxide (NO 2 ), 73 μg/m 3 nitrogen oxides (NO x ), 14 μg/m 3 particulate matter with aerodynamic diameter <2.5 μm (PM 2.5 ), 23 μg/ m 3 particulate matter with aerodynamic diameter <10 μm (PM 10 ), and 32 μg/m 3 ozone (O 3 ). Average daytime (L Aeq,16hr ) and night-time (L night ) road traffic A-weighted noise levels were 58 dB and 53 dB respectively. Interquartile range increases in NO 2 , NO x , PM 2.5, PM 10 , and source specific PM 2.5 from traffic exhaust (PM 2.5 traffic exhaust ) and traffic non-exhaust (brake or tyre wear and resuspension) (PM 2.5 traffic non-exhaust ) were associated with 2% to 6% increased odds of term LBW, and 1% to 3% increased odds of term SGA. Air pollutant associations were robust to adjustment for road traffic noise. Trends of decreasing birth weight across increasing road traffic noise categories were observed, but were strongly attenuated when adjusted for primary traffic related air pollutants. Only PM 2.5 traffic exhaust and PM 2.5 were consistently associated with increased risk of term LBW after adjustment for each of the other air pollutants. It was estimated that 3% of term LBW cases in London are directly attributable to residential exposure to PM 2.5 >13.8 μg/m 3 during pregnancy. The findings suggest that air pollution from road traffic in London is adversely affecting fetal growth. The results suggest little evidence for an independent exposure-response effect of traffic related noise on birth weight outcomes.