Measuring the Environment and Exposome

Hui Hu Ph.D.

Department of Epidemiology

College of Public Health and Health Professions & College of Medicine

December 5, 2018

Measuring the Environment



Measuring the Exposome

Measuring the Environment


  • Natural environment includes all things naturally occurring on Earth
  • Physical, chemical, and biological components or contaminants in soil, air, water
  • Exposure to these agents are usually unknown or unsensed by the exposed individual
    -  the exposure cannot be recalled or recorded by the study subject
    -  it can only be documented through measurements in the environment

Two types of errors

  • Error variance (lack of precision) in the measure
    -  individual levels of exposure are over- or underestimated but the mean exposure of a group is correct
    -  usually leads to attenuation of the observed relationship of the exposure to outcome
  • Systematically incorrect estimation of exposure levels
    -  the mean exposure level for a group is over or under-estimated
    -  can lead to over- or underestimation of the risk at a fixed level of exposure

Influential factors of environmental measurements

  • Objectivity: depends on two steps
    -  the sampling of the environment
    -  the measurement procedure
  • In some case, both steps can be highly objective (e.g. measure the concentration of gaseous air pollution using personal sampler)
  • Often only the second step is objective, while identification and selection of the sample involves substantial subjectivity
    -  rely on information provided by the study subject to identify environmental materials to be sampled (e.g. use residential history to assess air pollution exposure)

Influential factors of environmental measurements (continued)

  • Individualized:
    -  can be achieved only when personal sampling is practicable over repeated and extended periods of time
  • In many cases, it can only be approximated by average measurements over time or groups of persons
    -  estimating air pollution exposure using a stationary air monitor
  • Errors in obtaining the original specimens from the environment and in sub-sampling for lab analysis may be the main source of error in environmental measurements

Influential factors of environmental measurements (continued)

  • Analytic specificity and sensitivity
  • If the exposure is a mixture of chemical, measurements specific for a single chemical may be inappropriate if
    -  that chemical is not the one responsible for the biological effect under study,
    -  and its concentration is not highly correlated with that of the active agent or with the total activity in the mixture
  • Cost and feasibility of use on a large scale

Sampling and measuring present exposures

  • Two main purposes:

    -  to be used as such in prospective cohort studies aimed at relating present exposure to future disease occurrence

    -  to be entered as one element in the process of estimating past exposure within cross-sectional, retrospective cohort, or case-control studies

Methods of environmental sampling

  • Measurement at fixed points within the environment and inferring individual exposure from the concentration of contaminants measured in the parts of the environment covered by each sampler
  • Measurement of the immediate and continually changing environment of individual subjects by some form of personal sampling
    -  this is preferred since it takes into account the subject's position in the environment, and concentrations of the environmental agent there, and behaviors which may modify exposure in particular circumstances

Selecting subjects for environmental sampling

  • Continuous sampling of the environment of each subject throughout all periods of exposure relevant to the study
    -  relatively simple in some settings: e.g. film-badge monitoring of exposure to ionizing radiaiton
    -  usually impracticable and unnecessary
  • Sampling both subjects and exposure time
  • Two main approaches:
    -  random selection of subjects and then grouping those appearing to share common levels of exposure to the environmental agent
    -  pre-definition of homogeneous strata with respect to the exposure and sampling randomly within the strata

Sampling and measuring past exposures

  • Measurements of present exposure in the environment are applicable to cross-sectional, case-control, or retrospective cohort studies only when it can be reasonably assumed that measurements in the present environment are highly correlated with the past
  • Records of past measurements of relevant exposures can be sought
  • Three scenarios:
    -  complete data
    -  incomplete data
    -  data unavailable
  • It is practically impossible to monitor all agents present in the environment

Complete past measurements

  • Usually, the measurements on natural environment available were collected to check compliance with regulatory standards
  • Important to ascertain the purpose of the measurements and the frame within which they were collected
    -  they determine the way in which the samples were selected and the way they reflect the pattern of exposure of the study subjects
  • Example:
    -  measurements of only a fraction of the mixture of interest
    -  measurements only in "maximum risk employees"

Incomplete past measurements

  • Incomplete:
    -  lack of spatiotemporal coverage
    -  measured by sub-optimal methods
    -  measured only some correlate of the exposure of interest
  • Methods to address these problems:
    -  spatiotemporal interpolation
    -  calibration and conversion
    -  predictive models


Phenotypes are a function of inherited and environmental factors

P = G + E




T2DM, Cancer, LBW, PTD


Infectious agents





  • There are exquisite tools that have been developed to sequence the human genome and to interrogate individual susceptibility through genome-wide association studies (GWAS)
  • Thousands of GWAS


  • However, there was a lack of comparable tools in relation to exposure assessment
    -  almost uniquely focused on single exposure-health effect relationships and no global view of how various types of exposures co-exist and jointly affect health
  • A similar platform for discovery should exist for E
    - why?

  • Heritability: the range of phenotypic variability attributed to genetic variability in a population
    -  indicator of the proportion of phenotypic differences attributed to G
H^2={{\sigma^2_G}\over {\sigma^2_P}}
H2=σG2σP2H^2={{\sigma^2_G}\over {\sigma^2_P}}


  • To draw attention to the critical need for more complete environmental exposure assessment
    -  environment is defined in this context as 'non-genetic' factors
    -  exposome complements the genome
  • The exposome is composed of every exposure to which an individual is subjected from conception to death
    -  the nature of those exposures
    -  their changes over time

Three domains of exposome

  • Three broad domains of non-genetic exposures:
    -  internal
    -  specific external
    -  general external

Source: Wild CP. The exposome: from concept to utility. International journal of epidemiology. 2012 Feb 1;41(1):24-32.

Three domains of exposome (continued)

  • There is overlap in the three domains
    -  physical activity can either be internal or specific external
  • The domains can also be considered as intertwined
    -  the internal may at least partially be a response to the external
  • Measures in one domain or another may reflect to differing degrees one component of the exposome:
    -  the urban environment (general external)
    -  air pollution (specific external)
    -  inflammation (internal)

Source: Wild CP. The exposome: from concept to utility. International journal of epidemiology. 2012 Feb 1;41(1):24-32.

Exposome vs. traditional epidemiologic approaches

  • Expanded and dynamic exposure assessment across multiple domains
  • Integration of data on exposure and response across multiple scales of variation
  • Use of the resulting high dimensional information on multiple exposure-response relationships for data-driven discovery


  • One of the most challenging features of exposome's characterization
  • At any given time, an individual will have a particular profile of exposures
  • To fully characterize an individual's exposome:
    -  sequential measures that spanned a lifetime
    -  a small number of measures that captured exposure over a series of extended periods

Source: Wild CP. The exposome: from concept to utility. International journal of epidemiology. 2012 Feb 1;41(1):24-32.

  • Periods of only gradual change in the exposure profile
  • Periods with dramatic changes in specific components
    -  e.g. change in occupation
  • Periods when there is a radical change in global exposure profile in a short period
    -  e.g. one is born or migrates

Measuring the Exposome

Approaches and tools to measure the exposome

Source: Wild CP. The exposome: from concept to utility. International journal of epidemiology. 2012 Feb 1;41(1):24-32.

Bottom-up vs. top-down


  • Genome sequencing has been closely followed by analogous tools to characterize downstream biological event:
    -  RNA expression (transcriptomics)
    -  proteins (proteomics)
    -  metabolites (metabolomics)
    -  epigenomics
  • Each of the omics provides information on thousands of individual component parts, and can be linked to disease outcomes in analogous fashion to GWAS

  • Some of the limitations to use these approaches in exposome:
    -  the 'top-down' approach lacks a crucial link between exposure to a given agent and the molecular profiles observed
    -  this is required to identify risk factors and to implement prevention strategies

Other methods

  • Additional challenges:
    -  under-representation of individuals with lower SES
    -  ensure that the inclusion of new technologies helped reverse rather than exacerbate these problems of recruitment

Timing of measurement

  • To obtain the totality of the exposome:
    -  methods that integrate the fluctuations over time
    -  a series of snapshot measures at specific times in the lifespan of an individual
  • It is unlikely to capture long-term past exposures in a comprehensive way
  • Several key stages proposed for the snapshot approach:
    -  gestation
    -  early childhood
    -  puberty
    -  reproductive years

The Health and Environment-wide Associations based on Large population Surveys (HEALS)

Aims to develop an integrated methodology for compiling and organizing exposure data across multiple domains, and using these factors to predict health outcomes

Human Early-Life Exposome (HELIX) Project

Aim to implement tools and methods to characterize early-life exposure to a wide range of chemical and physical environmental factors and associate with data on major child health outcomes
-  early-life exposome approach


A good online tutorial for exposome analytics