Design and subjects
. The present cross-sectional study includes 10-year-old children from the Environment and Childhood Asthma (ECA) study in Oslo, Norway, which is described in detail elsewhere (Lødrup Carlsen 2002
). Briefly, this birth cohort included 3,754 healthy infants born in Oslo during a 15-month period beginning 1 January 1992. The ECA study design involved clinical investigations of two main study populations within the birth cohort. Lung function was obtained in 802 infants shortly after birth (Lødrup Carlsen et al. 1994
). At 2 years of age, a nested case–control study aimed to include all children from the initial birth cohort who had recurrent (> 1) or persistent (> 4 weeks) doctor-confirmed bronchial obstruction, along with age-matched controls (Lødrup Carlsen 2002
). Children with lung function measured at birth or who had participated in the case–control study at 2 years of age were invited to a 10-year follow-up; 1,019 (84%) children participated in the investigation [see Supplemental Material, Figure S1 (http://dx.doi.org/10.1289/ehp.1205256
Of the 1,019 children who participated in the 10-year follow-up, funds were available to analyze urine specimens for 623 children. Subjects who had current asthma at 10 years were preferentially sampled [see Supplemental Material, Figure S1 (http://dx.doi.org/10.1289/ehp.1205256
)], yielding a higher prevalence of current asthma (21%) than in the 396 children who participated in the ECA 10-year follow-up but were not included in the present study (7%) (see Supplemental Material, Table S1). The 623 children in the present study had a slightly lower median age (10.7 vs. 10.8 years), a higher prevalence of atopic eczema (23% vs. 18%), a lower prevalence of allergic sensitization [a positive skin prick test or elevated serum-specific IgE (≥ 0.35 kU/L) for at least one of the following allergens tested: house dust mites (Dematophagoides pteronyssinus
and D. farinae
), cat, dog, rabbit, German cockroach, birch, timothy, and mugworth pollen, Cladosporium, Alternaria
, egg white, milk, peanut and codfish; 35% vs. 43%] and were less likely to be firstborn (49% vs. 56%) than the 396 excluded children. The included children were comparable to the remainder of the original study cohort at birth with regard to most characteristics, but were less likely to be first-born (49% vs. 56%) and to be in a middle-income family (see Supplemental Material, Table S2).
The 10-year follow-up of the ECA study was approved by the Regional Committee for Medical Research Ethics and the Norwegian Data Inspectorate, and the biobank was registered in the Norwegian Biobank Registry (Oslo). The parents gave written informed consent for their child’s participation in the study. The involvement of the Centers for Disease Control and Prevention (CDC) laboratory was determined not to constitute engagement in human subjects research.
Clinical examination and outcome definition. The 10-year follow-up was performed between September 2001 and December 2004, and included a detailed parental structured interview, blood tests, skin prick tests (SPT) for allergic sensitization, spirometry, first morning urine sampling, and clinical examinations.
Current asthma (Lødrup Carlsen et al. 2006
) was defined by a history of asthma reported by the parent plus at least one of the following: reported dyspnea, chest tightness, and/or wheezing during the previous 12 months; reported use of asthma medication during the previous 12 months; or a positive exercise challenge test (a ≥ 10% decrease in forced expiratory volume in 1 sec compared with baseline, performed 2–7 days after the first visit).
Urine collection and analyses
. The participants’ first morning voids, collected at home in commercial polypropylene specimen collection containers, were aliquoted and frozen at –80ºC in 2 mL polypropylene tubes until they were shipped on dry ice to the CDC (Atlanta, GA, USA) for analysis. Three low-MWP metabolites [molecular weight < 250 Da; MEP, mono-n
-butyl phthalate (MnBP), and monoisobutyl phthalate (MiBP)] and eight high-MWP metabolites [molecular weight > 250 Da; monobenzyl phthalate (MBzP), mono(3-carboxypropyl) phthalate (MCPP), monocarboxyoctyl phthalate (MCOP), and monocarboxynonyl phthalate (MCNP), including four metabolites from DEHP: MEHP, mono(2-ethyl-5-oxohexyl) phthalate (MEOHP), mono(2-ethyl-5-hydroxyhexyl) phthalate (MEHHP), and mono(2-ethyl-5-carboxypentyl) phthalate (MECPP)], were measured in the urine by online solid phase extraction coupled with high-performance liquid chromatography with isotope-dilution tandem mass spectrometry using a modification of the method reported by Silva et al. (2007)
. We multiplied the reported MEP and MBzP concentrations by 0.66 and 0.72, respectively, to correct for the inadequate purity of the analytic standards used. The lower-than-expected purity of the standards does not affect the performance of the analytical methods (CDC 2012
). After analyses, the urine samples were shipped on dry ice to the Norwegian Institute of Public Health. Urine specific gravity (SG) was measured by a handheld refractometer (Atago Urine Specific Gravity Refractometer PAL 10-S; ATAGO USA, Inc., Bellevue, WA, USA) after thawing and thorough mixing of the sample with a vortex shaker. The refractometer was calibrated with distilled water between each measurement.
Statistical analyses. SG-corrected urinary concentrations of phthalate metabolites were compared between girls and boys by analysis of variance (ANOVA). SG-corrected metabolite concentrations were calculated by the formula Pc = P[(1.024-1)/(SG-1)], where Pc is the SG-corrected phthalate metabolite concentration (micrograms per liter), P is the observed phthalate metabolite concentration, 1.024 is the median SG value in the study population, and SG is the specific gravity of the individual urine sample. For phthalate metabolite concentrations below the limit of detection (LOD), we used an imputed value equal to LOD/√–2. Correlations between the individual phthalate metabolites were evaluated using Spearman rank correlation coefficients (rS).
For current asthma, we fit logistic regression models by quartile concentration categories of the metabolites (lowest concentration category as reference) and a second model with phthalate metabolites as continuous variables (log10-transformed). SG was used as a covariate in the models. We modeled each of the individual phthalate metabolites separately, except for the DEHP metabolites (MEHP, MEOHP, MEHHP, and MECCP), which were combined into their micromolar sum (sum of the individual metabolites concentrations in micrograms per liter divided by the molecular weight of the metabolites) (ΣDEHP). We fitted additional models for the micromolar sums of high-molecular weight (Σhigh-MWP) and low-molecular weight (Σlow-MWP) metabolites. To test for trend across quartiles, we fitted the median quartile concentration values as a continuous variable in the logistic regression model.
Covariates were selected based on a directed acyclic graph [see Supplemental Material, Figure S2 (http://dx.doi.org/10.1289/ehp.1205256
)] that included the following variables: sex, body mass index (BMI; age and sex adjusted), allergic sensitization in the child, parental smoking at home [between the school age of the child (6–7 years) and the 10-year follow-up], parental asthma (at child’s birth), maternal education (at child’s birth), and household income (at the 10-year follow-up). The minimal adjustment set for a total effect of phthalates on current asthma was sex, parental asthma, and household income. To examine the role of allergic sensitization in the association between phthalate metabolites and asthma, we fitted additional models after stratification by allergic sensitization. Finally, we tested for modification effects by sex, by adding interaction terms to the models. p
-Values ≤ 0.05 were considered statistically significant. All statistical analyses were performed with Statistical Package for Social Sciences (SPSS version 19.0; IBM Inc., Chicago, IL, USA).