, 2000). For chemicals with short half-lives, however, the interval between the relevant exposure and disease development is often difficult to assess. Study design – along with exposure misclassification discussed later in
this paper – are the most critical and underexplored aspects of biomonitoring studies of short-lived chemicals. Establishing temporality is much more difficult in a “prevalence” study compared to an “incidence” study, which makes it challenging to draw conclusions about causal associations. A typical prevalence study relies on cross-sectional INK 128 concentration design, which ascertains the exposure and disease information simultaneously (Rothman and Greenland, 1998). When research is focused on short-lived chemicals, many case–control studies – even if they use incident cases – are difficult to interpret because the biomarker levels reflect recent exposures that typically follow rather than precede disease onset. The notable exception is a study that uses samples collected and stored for future use, as is done in nested case–control or case–cohort studies (Gordis, 2008). In a recent review of the epidemiology literature on phthalate metabolites (Goodman et al., 2014) and their association with obesity, diabetes, and cardiovascular disease, most of the studies were cross-sectional in design. The study results this website were inconsistent across
outcomes and lack of temporality was identified as a key limiting factor in the Galactosylceramidase ability to discern relationships between prior exposures to phthalate metabolites and consequent health outcomes. Tier 1 studies are incidence studies that involve a follow-up time period or a longitudinal analysis of repeated measures and allow the establishment of both the time order and the relevant interval between the exposure and the outcome (Table 1). A Tier 2 study would include incidence studies in which
exposure preceded the outcome, but the specific relevant windows of exposure are not considered. The least informative (Tier 3) studies are those that examine the association between current exposure (e.g., blood level of a chemical) and frequently measured outcomes (e.g. BMI) that are likely associated with chronic rather than acute exposures. (Note that this evaluative criterion is not applicable to studies focused on exposure only, such as those examining temporal or spatial relationships within or across populations.) For many short-lived chemicals, there can be large intra-individual temporal variability; attempting to find associations between one measure of such a chemical with disease is not supportable. Differences in biomonitored levels of short-lived chemicals due to changes in an individual’s diet, health, product use, activity and/or location are expected (Pleil and Sobus, 2013). As noted by Meeker et al.