Variability of urinary concentrations of non-persistent chemicals in pregnant women and school-aged children

Applying the exposome concept in birth cohort research: a review of statistical approaches

“…While single substitution could be acceptable when the proportion of values below the LOD is low (e.g., < 5%), this method introduces bias in the results as the proportion of values below the LOD increases [26]. It is therefore recommended that values below the LOD are imputed using imputation approach for left-censored missing data (e.g., using the imputeLOD function available in the rexposome R package) [27,28]. However, despite the good performance of the imputation methods, exposures with a high proportion of values below the LOD (e.g., > 80%) should be either not used or dichotomized into detected/undetected.…”

“…However, this cannot be computed if no repeated biospecimen are collected, at least in a subset of the study population and the literature cannot always provide relevant intraclass correlation coefficient estimates since these need to be assessed from comparable studies, typically in terms of study population and of biospecimen collection time points. Intraclass correlation coefficient for some non-persistent chemicals have been recently reported [27]. Previous studies have described methods that jointly correct for measurement error and perform variable selection, but they are mostly complex to implement, not available in standard statistical software and not applicable to all types of regression models or to all settings [29][30][31].…”

“…Large within-subject temporal variability is a common issue for non-persistent chemical contaminants as they can have short half-life in the human body. As a result, for many non-persistent chemicals, a few dozen samples seem to be required to accurately assess exposure over periods encompassing several trimesters or months [19,27,42]. A recent study, characterizing the personal external exposome, including air pollution, traffic-related noise, natural outdoor environments, ultraviolet radiation and levels of physical activity, of pregnant women and children in eight European cities, suggested that the assessment of these personal exposures requires monitoring from one day to more than one year, depending on exposure due to high variability between and within cities and participants [43].…”

A Scoping Review on the Characteristics of Human Exposome Studies

“…Approximately a third of the retained exposome studies had a prospective longitudinal design (n = 30, 37%), of which 26 were based on nine large population-based cohorts; the rest four longitudinal studies were relatively small in size, including between 6 and 378 subjects [46][47][48][49]. Out of the 26 studies, eleven were published as part of the Human Early-Life Exposome (HELIX) project using data from six European birth cohorts ( [50][51][52][53][54][55][56]), and among these studies, four belonged to the Spanish INMA (INfancia y Medio Ambiente) [57][58][59][60][61]. Four studies (of the 26) were published as part of the Avon Longitudinal Study of Parents and Children (ALSPAC), a 20-year population-based cohort designed to determine the influence of environmental (physical and psychological) and genetic factors on the health status and development of the offspring.…”

“…On the other hand, three studies (4%) explored the variability of the metabolomic profile through time: between sample collection time-points; pregnancy trimesters; or lifespan in women over 40 years old [51,68,85]. Four studies (5%) investigated the differentiation of metabolomics profiles by the following health end points/outcomes: oxidative stress, sarcoidosis, Crohn's disease, premature death, fetal growth, and blood pressure changes during pregnancy [56,83,96,102].…”