As expected from the ubiquitous use of phthalates in modern societies, our data suggest widespread human exposure among this adult study population. The high detection frequency and concentrations were within the ranges reported for the adult U.S. general population from the National Health and Nutrition Examination Survey (NHANES) 2005–2006 (CDC 2010
Six previous studies have evaluated the temporal variability of phthalate metabolites in various populations over periods ranging from days to months (Adibi et al. 2008
; Fromme et al. 2007
; Hauser et al. 2004
; Hoppin et al. 2002
; Peck et al. 2010
; Teitelbaum et al. 2008
). In three of these studies, researchers assessed the agreement of phthalate metabolite concentrations by using first morning urine samples collected from 46 African-American women on 2 consecutive days (Hoppin et al. 2002
), from 50 German men and women 14–60 years of age during 8 consecutive days (Fromme et al. 2007
), and from 25 Hmong women who provided up to three samples over approximately 30 days (Peck et al. 2010
). In the other studies, researchers evaluated the variability of phthalate metabolite concentrations in multiple spot urine samples collected from 28 Dominican and African-American women who provided two to four samples over 6 weeks during their last trimester of pregnancy (Adibi et al. 2008
), from 11 men who provided nine samples each over 3 months (Hauser et al. 2004
), and from 35 Hispanic and African-American children 6–10 years of age who collected two to seven samples over 6 months (Teitelbaum et al. 2008
). Except for one study (Fromme et al. 2007
), all others assessed the variability of MEP, although two older reports (Hauser et al. 2004
; Hoppin et al. 2002
) did not assess the variability of MEHHP and other DEHP oxidative metabolites.
In agreement with the ICCs reported previously for MEHHP urinary concentrations of first morning voids over 8 days to ~ 1 month (Fromme et al. 2007
; Peck et al. 2010
) and of spot samples collected within 6 weeks to ~ 6 months (Adibi et al. 2008
; Teitelbaum et al. 2008
), we found that between-person MEHHP creatinine-corrected concentrations for our study population varied considerably over 7 consecutive days (ICC = 0.25 for first morning voids; ICC = 0.17 for spot samples). By contrast, we found a low variability of between-person creatinine-corrected concentrations of MEP during the same time period (ICC = 0.91 for first morning voids; ICC = 0.77 for spot samples). Other studies have also reported moderate (ICC = ~ 0.6) reproducibility in MEP urinary measures (Hauser et al. 2004
; Hoppin et al. 2002
; Peck et al. 2010
In this study, the largest contribution of the total variance of MEP urinary concentrations in spot samples was the between-person variability (77%). Throughout the day, the average person’s variance was also considerable (21%), but the average person’s between-day variance was rather low (2%). Similarly, the largest percentage of total variance in MEP concentrations from first morning and 24-hr voids was also the variation between each person (91% and 94%, respectively).
DEP exposure is largely associated with the use of personal care products (Api 2001
; Berman et al. 2009
; Duty et al. 2005
; Houlihan et al. 2002
; Hubinger and Havery 2006
; Koo and Lee 2004
; Sathyanarayana et al. 2008
; Schettler 2006
). The large between-person variability of MEP urinary concentrations we observed among this group of adults is likely related to the fact that different people use different types and combinations of personal care products. On the other hand, the fact that the persons examined had a large variation in MEP urinary concentrations throughout a given day, but very small variation between days, may be related to their regular use of personal care products. We speculate that people typically use the same personal care products at similar times in their daily routines and that individuals also tend to apply personal care products in similar amounts and frequency every day. Furthermore, the regular use of personal care products at similar times every day and the short DEP elimination t1/2
could result in MEP being excreted every day at similarly spaced times. This behavioral use of DEP-containing personal care products may also explain the appearance of a cyclic pattern in MEP urinary concentration in the persons with the largest concentrations of MEP during the study week. Of interest, the cyclic pattern was particularly evident during the work week but seemed to change for many participants over the weekend.
Unlike MEP, the largest variation of MEHHP urinary concentrations in spot samples was related to the variation of each person throughout the day (51%). The within-person variability between days was also considerable (32%) and about twice the variation attributed to differences between persons (17%). Likewise, the largest contributor to the total variance of MEHHP concentrations in first morning and 24-hr urine voids was individual variability from day to day (75% and 69%, respectively). We obtained similar results for MEHP, the DEHP hydrolytic metabolite, even though MEHHP elimination t1/2
is 3–4 times longer than of MEHP, suggesting that the main factors affecting the observed variance for MEHHP and MEHP concentrations are similar. Exposure to DEHP, the MEHHP precursor in the body, is largely associated with the consumption of food (Kavlock et al. 2006
). Not only do diets vary from person to person, but an individual’s food consumption typically changes from day to day. Consistent with this, we did not observe clear daily patterns in MEHHP urinary concentrations for most participants.
Our findings also suggest that, regardless of the type of sample collected (i.e., spot, first morning, and 24-hr voids), when diet is the likely main source of exposure (i.e., DEHP), interday variability is a main contributor to the total variance. By contrast, when routine daily use of a product is the main exposure source (i.e., DEP), interperson variability appears to be the main contributor to the total variance. However, age will have a strong impact in relation to exposure to environmental chemicals, including phthalates, because behavior and diet, among other factors, are likely contributors to exposure to these compounds. For example, for young children, particularly infants, diet may not be as diversified as it is for adults. The extent and patterns of use of personal care products among children and adults are also expected to differ. Therefore, some of the findings we report for this group of adults may not apply to children and other age groups. Furthermore, the number of study participants examined was rather small, although the reported MEP and MEHHP urinary concentrations fell within the NHANES reference ranges. For the above reasons, we recommend caution in the generalization of our findings to other populations.
Twenty-four hour urine specimens do not require a correction for the urine dilution, which is important because no consensus exists on the best method for conducting such adjustment (Adibi et al. 2008
; Barr et al. 2005
; Pearson et al. 2009
). On the other hand, our findings suggest that collecting 24-hr samples for only 1 day could benefit studies designed to evaluate compounds to which people are mostly exposed through routine use of personal care products. Unfortunately, for many chemicals, the contribution to the total exposure from all potential sources is either variable or unknown. In addition, epidemiologic studies often evaluate exposure to a wide range of compounds (and their corresponding exposure sources). As a result, collecting 24-hr voids may not necessarily eliminate the potential for exposure misclassification, at least for some of the compounds examined. Therefore, when the population is sufficiently large, the spot-sampling approach may provide enough statistical power to adequately categorize exposure, particularly when samples are collected on multiple days.
One of the most important findings of this work is to show that, for a given person, the urinary concentrations of phthalate metabolites can change considerably throughout the day. Others have observed similar intraday variability in the urinary concentrations of other nonpersistent chemicals, such as PAH metabolites (Li et al. 2010
). More important, even for the two metabolites we evaluated, the intraday changes went in opposite directions. For example, we found that the lowest MEP urinary concentrations, but the highest MEHHP concentrations, occurred in the evening. These findings suggest that sampling strategy should be one critical factor when designing epidemiologic studies that include biomonitoring measures of urine specimens. Very often these specimens are analyzed for more than one class of environmental chemicals. Therefore, when multiple collections of spot urine samples over a period of days to weeks or months are logistically and economically possible, specimens should be collected at different times of the day. Our data for MEP and MEHHP suggest that this approach would maximize the suitability of the urinary concentrations of the various target biomarkers to reflect temporal exposure to nonpersistent chemicals. However, when and how often urine samples are collected will depend not only on how reproducible the urinary concentrations are (i.e., relatively high ICC), but also on the target population, aims of the study, major route of uptake of the parent phthalate, and excretion t1/2
of its metabolites. For DEP, sampling around midday on any given day may be advantageous if exposure occurs mainly through the use of personal care products and if these products are applied in the morning, because MEP excretion t1/2
is 3–4 hr and peak excretion would be expected to occur around midday. Whether multiple sampling is needed for exposure assessment in specific situations (e.g., during pregnancy) will depend mainly on the intraperson variability at the sampling time (e.g., noon) and throughout the study period (e.g., 1 week). For DEHP, because of the strong influence of diet, the daily intraperson variability may be as high as the intraperson variability at one specific time of the day throughout the course of the study. When intraperson variability is unavoidable and highly independent of the sampling time, two potential approaches for conducting exposure assessment are as follows: a
) use the mean or median urinary concentrations of all of the samples collected over a certain time period if multiple collections per person are possible, or b
) if only one spot sample per person is available, use each individual concentration and provide estimates of upper and lower confidence intervals (CIs) based on results from all participants. For the latter, our findings might serve as a basis for setting such fixed upper and lower CIs of exposure, in particular, for epidemiologic studies where recruitment of participants has been completed and multiple sampling for exposure assessment is no longer possible.
Among participants of NHANES 1999–2000, variations have been reported in the distributions of urinary concentrations of MEP and other phthalate metabolites, depending on the time of day of sampling (Silva et al. 2004
). In addition to this variability, we found that MEP urinary concentrations also differed by age and race/ethnicity (Silva et al. 2004
). Although the nature of the exposure to phthalates and the short t1/2
of the phthalates will affect the urinary concentrations of phthalate metabolites on an individual basis, on a population basis, the range of concentrations observed in our study may represent an average exposure scenario. For example, MEP concentrations in the upper percentiles resulting from the collection of urine soon after DEP-related activity of an individual will likely be offset by a urinary concentration in the lower percentiles originating from another person who provided a sample shortly before conducting the same activity. In this study, the considerable variation in concentrations of urinary biomarkers of DEP and DEHP suggests considerable variability in exposure among adults to these two phthalates selected to represent two main daily activities: use of personal care products (dermal exposure to DEP) and diet (ingestion exposure to DEHP). We hypothesize that the patterns of exposure variability observed for DEP and DEHP will encompass those of other phthalates such as dibutyl and benzylbutyl phthalates, which do not have a clearly identified predominant pathway of exposure for the average adult person. However, additional research is needed to assess the variability in exposure to other phthalates and among populations that encompass different lifestyles and life span stages.