Table 2 Recommendations for methods, evaluations of collaboration with existing data, and areas of interest for future work including new data/new methods by exposure topic

Outdoor air pollution

·Exposure modeling is currently the state-of the art method

·Within the ESCAPE project, a standardized exposure assessment is being added to a number of birth cohort studies and will soon be linked to existing health data in these cohorts; pooled analyses will be performed for a number of health endpoints

·Assessment of long-term exposure to ultrafine particles, which are currently not being assessed within cohort studies.

·Few studies so far compared land-use regression models with dispersion models; results are inconsistent

·Residential mobility, time-activity pattern, and exposure at non-residential addresses should be evaluated in exposure assessment

·Assessment of long-term validity (i.e. stability) of land-use regression models which are based on one measurement campaign

·There is currently little validation of modeled exposures against personal exposure measurements

Water contamination

·Best method for exposure assessment is to combine information of water-related behaviors obtained by questionnaire with newly or routinely collected water contaminant measurements

·Further opportunity to study exposure to water pollutants in cohorts without water exposure assessment; routinely collected water pollutants are often available

·Assessment of exposure to substances such as pharmaceuticals, Perfluorooctanesulfonic acid (PFOS)/Perfluorooctanoic Acid (PFOA), and (other) endocrine disrupters

·Validation of questionnaires against biomonitoring since it has been conducted in only a few subjects

·Consideration of source of water as an exposure indicator (ground water versus surface water)

·Assessment of variability and measurement error of questionnaires by repeated assessments

·Data pooling currently being done in the HIWATE project

·Access to publicly collected data should be increased and European databases should be made available to researchers

Allergens & microbial agents

·Measurements in house dust or air samples are recommended

·Meta-analyses in the framework of the GA2LEN (allergens) and AIRALLERG projects (allergens and biocontaminants) and ENRIECO case study on mold and dampness

·Use of newly developed analysis techniques such as molecular methods or DNA fingerprinting

·Use of questionnaires is inexpensive, but questionnaires were found to have a low sensitivity

·Exposure at non-residential locations and timing of exposure should be taken into account

Metals

·Human biological monitoring is the state of the art method for estimation of total dose.

·Data pooling and/or meta-analysis of the data available in the European birth cohorts can overcome this problem if conversion models can be developed to transfer between different biological media (hair, cord blood, urine, etc.).

·Validation of questionnaire data against human biological monitoring is needed.

·Inductively coupled plasma mass spectrometry (ICP-MS) is more sensitive and faster than Atomic absorption spectroscopy (AAS)

·In general labs are using well standardized protocols. It is recommended to validate the analytical technique in each lab including a sample with known concentration of metal/s every X number of samples. This will be useful to validate both the pre-treatment and the analytic process.

·Some studies compared Mercury levels in biological samples with fish/shellfish consumption assessed through validated food frequency questionnaires with encouraging results. More studies are needed to confirm these findings. For other metals, validation of other exposure assessment methods must be further explored.

Pesticides

·There are multiple pesticides and multiple pathways of exposure conducing to varying exposure assessment

·Since few cohorts assessed exposure to pesticides there is a large scope of doing more work on pesticide exposure within the European birth cohorts, particularly by analyzing available biological samples

·Include validation studies in exposure assessment

·Biomonitoring hardly feasible for large cohorts, but recommended on sub-populations for validation purposes and identification of main exposure sources

·Use of Geographic Information System (GIS) technologies with existing European data on soil occupancy and satellite imagings/maps of crops to assess bystander exposure due to agricultural activities

·Assessment of time-activity pattern and exposure at non-residential locations

·There is a need for harmonization of exposure assessment by biomonitoring (choice of molecules of interest, biological matrices, sampling and storage conditions, chemical analyses controlled by international comparison programs, etc.) between studies

·Validation of questionnaires needed

·Inclusion of exposure at non-residential locations may improve exposure assessment

Persistent organic pollutants

·High-performance liquid chromatography (HPLC) derived methods are the state of the art for measurements of POPs

·The possibility to perform a pooled or meta-analysis on the association between exposure to POPs and birth outcomes in the European birth cohorts have been evaluated within a case study that is part of ENRIECO and continued in CHICOS (http://www.chicosproject.eu).

·There is high degree of variability between studies in study design including timing of sample collection and collection medium. Therefore additional data collection according to a standardized protocol may be needed, especially if the outcomes of interest are hypothesized to be related to exposure at specific time windows during fetal life or early childhood.

·All analyzing laboratories should participate in inter-laboratory calibration tests.

·Data pooling is possible for polychlorinated biphenyl (PCB), and dichlorodiphenyltrichloroethane (DDT). For other POPs there is little data or too much heterogeneity with regard to the sampling media or the timing of exposure assessment.

·Especially for detecting POPs in low concentrations in small volumes equipment with a high sensitivity is needed.

·Conversion factors needed to be developed to allow pooling of data.

·The persistence of organochlorines makes sample degradation a lesser problem as for other more readily degradable compounds. However, it is recommended to store samples at −80°C at least, if measurements are planned to be performed after several years.

Other chemical exposures (brominated flame retardants, perfluorinated compounds, phthalates and phenols)

Human biological monitoring is the state of the art method for estimation of total dose

There is currently little published data in the cohorts, but many measurements are ongoing and we recommend cohorts starting to work towards combined and comparison studies.

This is an emerging field and there is a rapidly growing expertise in the cohorts, which would benefit from continued communication and coordination.

·For non-persistent exposures with very short half-lives (phthalates and phenols), we recommend repeated measurements as standard practice.

·Conversion factors should be developed to transfer from concentrations in one medium/time point to another in order to compare/combine data from different cohorts.

·Very little is known about the effects of postnatal exposure to emerging chemicals, and therefore we recommend further evaluation of these new chemicals in children.

·Issues of contamination from storage materials and lab equipment, and storage conditions are of great importance and need to be addressed in depth. We recommend to closely follow published recommendations on sampling collection and packing, storage, and analysis (see COPHES website: http://www.eu-hbm.info/cophes).

·Active dialogue and partnership among the scientists representing the various disciplines would be essential for selecting new contaminants and setting prioritization for measurement in birth cohort studies.

·It is recommended to conduct a European evaluation of inter- and intra-laboratory variability.

·Validation of other exposure assessment methods such as questionnaires, occupational Job Exposure Matrices (JEMs), environmental measurements, and/or toxicokinetic models is needed

Radiation

Ionizing radiation

·Assessment of medical radiation exposures (X-ray, computer tomography (CT)-scan) by standardized questionnaires

·Existing data are not sufficient for pooled studies.

·Assessment of medical radiation exposure

·Assessment of occupational exposure by means of badge dose information or if not possible by questions on x-ray equipment and protective equipment used

·Evaluation of link with other EUROPEAN cohorts of children exposed to CT scans

·Assessment through job-exposure matrices difficult

UV

·ENRIECO has developed a set of core questions on sun exposure for different exposure-time windows that is recommended for use in cohort studies

·Existing data are not sufficient for pooled studies

·Inclusion of vitamin D and UV exposure related questions in cohort questionnaires.

Non-ionizing radiation

·Use of core set of standardized questions to assess mobile and cordless phone use

·Comparison studies with existing data comparing questionnaire data between cohorts

·Integration of standardized questions on use of mobile phones to facilitate future combined analyses of non-cancer effects

·Validation of questionnaires using information of other types of studies

·Collaborative efforts focusing on design of questions related to other RF-EMF sources (e.g. WiFi, new communication technologies, microwave ovens, baby phones)

·Coordination between cohorts in developing validated exposure models for RF and ELF-EMF

Smoking and second hand tobacco smoke

·Questionnaires are most suitable method for larger epidemiological studies and for assessment of long-term exposure

·Combined analyses on the effects of pre- and postnatal exposure to second hand tobacco smoke have been performed within a case study that is part of ENRIECO

·Large studies with close monitoring of second-hand tobacco smoke exposure before conception, during trimesters of pregnancy and during the first year of life to disentangle the role of exposure during different periods

·Relevance of the timing of the exposure (before conception; during pregnancy, infancy, childhood or later in life) is not clear. Large studies can enhance knowledge if exposure is assessed repeatedly during different time periods.

·Specific questions recommended for the different exposure periods

Noise

·Objective measures should be in accordance with the European Union’s Environmental Noise Directive (END) guidelines.

·Few European cohorts currently have data from objective noise assessments that could be combined

·Inclusion of objective and subjective exposure assessments

·Noise propagation modeling is recommended for large studies.

·Assessment of time-activity pattern

·Questionnaire- assessments of noise annoyance should be performed in addition to objective measures; standard scales can be recommended. Information on non-residential exposure, time-activity pattern, and insulation of buildings, window opening behavior and the position of bedroom in relation to source of noise should be included in exposure assessment.

Occupational exposures

·A number of Job Exposure Matrices (JEMs) have been built in Europe covering different periods of time and different types of exposures.

·Within the ENRIECO project, the possibility to perform pooled/meta-analyses of the association between adverse health outcomes and selected occupations of mothers and fathers during vulnerable periods has been explored; 14 cohorts are eligible for this analysis, 12 have already expressed their interest. A protocol for the analysis has been developed.

·For an adequate data collection on occupational exposures job title is not sufficient. In addition, one should collect description of task, type of industry, number of hours per week, and if possible name of company, existence of biomonitoring data. Free text should be kept in the data base for additional details.

·JEMs need to be validated against objective measures of exposure (work environment, biomarkers)

·A good training of coders should be organized for harmonization of occupation coding

·If JEMs are used, they should be country- and agent-specific since work environments differ between countries and time periods.

·Standardized questionnaires for physical load should be published

·To avoid any influence of birth outcome on the availability of occupational information and on its quality, we recommend that data should be collected before birth. This is mandatory when questionnaires on occupational exposures are used and optional for job title.

·The period of interest is around conception and each trimester - or at least one trimester of pregnancy (depending on the health outcome studied) for mothers, and before conception for fathers.