Temporal trends in phthalate exposures: findings from the National Health and Nutrition Examination Survey, 2001-2010

Abstract

Background: Phthalates are ubiquitous environmental contaminants. Because of potential adverse effects on human health, butylbenzyl phthalate [BBzP; metabolite, monobenzyl phthalate (MBzP)], di-n-butyl phthalate [DnBP; metabolite, mono-n-butyl phthalate (MnBP)], and di(2-ethylhexyl) phthalate (DEHP) are being replaced by substitutes including other phthalates; however, little is known about consequent trends in population-level exposures.

Objective: We examined temporal trends in urinary concentrations of phthalate metabolites in the general U.S. population and whether trends vary by sociodemographic characteristics.

Methods: We combined data on 11 phthalate metabolites for 11,071 participants from five cycles of the National Health and Nutrition Examination Survey (2001-2010). Percent changes and least square geometric means (LSGMs) were calculated from multivariate regression models.

Results: LSGM concentrations of monoethyl phthalate, MnBP, MBzP, and ΣDEHP metabolites decreased between 2001-2002 and 2009-2010 [percent change (95% CI): -42% (-49, -34); -17% (-23, -9); -32% (-39, -23) and -37% (-46, -26), respectively]. In contrast, LSGM concentrations of monoisobutyl phthalate, mono(3-carboxypropyl) phthalate (MCPP), monocarboxyoctyl phthalate, and monocarboxynonyl phthalate (MCNP) increased over the study period [percent change (95% CI): 206% (178, 236); 25% (8, 45); 149% (102, 207); and 15% (1, 30), respectively]. Trends varied by subpopulations for certain phthalates. For example, LSGM concentrations of ΣDEHP metabolites, MCPP, and MCNP were higher in children than adults, but the gap between groups narrowed over time (pinteraction < 0.01).

Conclusions: Exposure of the U.S. population to phthalates has changed in the last decade. Data gaps make it difficult to explain trends, but legislative activity and advocacy campaigns by nongovernmental organizations may play a role in changing trends.

Citation: Zota AZ, Calafat AM, Woodruff TJ. 2014. Temporal trends in phthalate exposures: findings from the National Health and Nutrition Examination Survey, 2001-2010. Environ Health Perspect 122:235-241; http://dx.doi.org/10.1289/ehp.1306681.

Figure 1

Association between phthalate metabolites and NHANES sampling cycle in the general U.S. population for (A) MEP (n = 11,071; parent phthalate = DEP) (p < 0.0001); (B) MnBP (n = 11,071; parent phthalate, DnBP; p < 0.0001); (C) MBzP (n = 11,071; parent phthalate, BBzP; p < 0.0001); and (D) ∑DEHP metabolites (n = 11,071; parent phthalate, DEHP; p < 0.0001). Models are adjusted for urinary creatinine. Data points represent LSGM and error bars represent 95% CIs. Corresponding numeric data are provided in Supplemental Material, Table S3. p-Value for the overall comparison between groups assessed by the Wald Test.

Figure 2

Association between phthalate metabolites and NHANES sampling cycle in the general U.S. population for (A) MiBP (n = 11,071; parent phthalate, DiBP; p < 0.0001); (B) MCPP (n = 11,071; parent phthalates, DnOP and a nonspecific metabolite of high-molecular-weight phthalates; p < 0.0001); (C) MCOP (n = 6,375; parent phthalate, DiNP; p < 0.0001); and (D) MCNP (n = 6,375; parent phthalate, DiDP; p = 0.004). Models are adjusted for urinary creatinine. Data points represent LSGM and error bars represent 95% CIs. Corresponding numeric data are provided in Supplemental Material, Table S3. p-Value for the overall comparison between groups assessed by the Wald Test.

Figure 3

Association between phthalate metabolites and NHANES sampling cycle in the general U.S. population by age for (A) MEP (p_interaction = 0.04), (B) ∑DEHP metabolites (p_interaction = 0.002), (C) MCPP (p_interaction = 0.0004), and (D) MCNP (p_interaction = 0.009). Estimates are from linear regression models of interactions between NHANES sampling cycles and age adjusted for urinary creatinine, sex, race/ethnicity, and PIR. Data points represent LSGM and error bars represent 95% CIs. Corresponding numeric data are provided in Supplemental Material, Table S4.

Figure 4

Association between phthalate metabolites and NHANES sampling cycle in the general U.S. population by sex for (A) MnBP (p_interaction = 0.03) and (B) ∑DEHP metabolites (p_interaction = 0.0001). Estimates are from linear regression models of interactions between NHANES sampling cycles and sex, adjusted for urinary creatinine, age (continuous), race/ethnicity, and PIR. Data points represent LSGM and error bars represent 95% CIs. Corresponding numeric data are provided in Supplemental Material, Table S5.

Figure 5

Association between phthalate metabolites and NHANES sampling cycle in the general U.S. population by race/ethnicity for ∑DEHP metabolites (A; p_interaction = 0.006) and by PIR for ∑DEHP metabolites (B; p_interaction = 0.01) and MCPP (C; p_interaction < 0.0001). Estimates in (A) are from linear regression models of interactions between NHANES sampling cycles and race/ethnicity adjusted for urinary creatinine, age (continuous), sex, and PIR. Estimates in (B) and (C) are from linear regression models of interactions between NHANES sampling cycles and PIR adjusted for urinary creatinine, age (continuous), sex, and race/ethnicity. Data points represent LSGM and error bars represent 95% CIs.