A biomarker validation study of prenatal chlorpyrifos exposure within an inner-city cohort during pregnancy

Abstract

Background: We previously documented significant decreases in chlorpyrifos concentrations in maternal personal and indoor air samples among pregnant African-American and Dominican women from New York City after the 2000-2001 restrictions on its residential use.

Objective: We undertook a biomarker validation study within the same cohort to evaluate trends over time in multiple biomarkers of prenatal chlorpyrifos exposure.

Methods: Subjects were enrolled between February 2001 and May 2004 (n = 102). We measured 3,5,6-trichloro-2-pyridinol (TCPy) in postpartum meconium (n = 83), repeat prenatal maternal spot urine samples (n = 253), and postnatal urine from the mothers (n = 73) and newborns (n = 59). We measured chlorpyrifos in postnatal maternal (n = 92) and umbilical cord (n = 65) blood.

Results: We did not detect TCPy in infant urine, but all other biomarkers showed a highly significant decrease in detection frequencies (chi2 = 7.8-34.0, p < or = 0.005) and mean ranks (p < or = 0.006, Kruskal-Wallis) among subjects enrolled in 2003-2004 compared with those enrolled in 2001-2002. Chlorpyrifos in maternal personal and indoor air declined 2- to 3-fold over the same period (p < 0.05). In 2001-2002 samples, TCPy levels in repeat prenatal urine were positively correlated (r = 0.23-0.56), but within-subject variability exceeded between-subject variability (intraclass correlation coefficient = 0.43); indoor air levels explained 19% of the variance in prenatal urine TCPy (p = 0.001). Meconium TCPy concentrations were positively correlated with chlorpyrifos in maternal and cord blood (r = 0.25-0.33, p < 0.05) and with TCPy in maternal urine (r = 0.31, p < 0.01).

Conclusions: Results suggest the biomarkers are reliable dosimeters to differentiate between groups with prenatal chlorpyrifos exposures varying by a factor of 2 or more and vividly illustrate the efficacy of residential restriction on chlorpyrifos to reduce the internal dose during pregnancy.

Keywords: biomarkers; chlorpyrifos; cord blood; indoor air; maternal blood; meconium; pregnancy; urine.

Figure 1

Percentage of subjects with TCPy concentrations in one or more of repeat maternal prenatal urine samples > LOD (A) and mean TCPy ranks adjusted for creatinine (B) by year of sample collection. Sample size: 2001, n = 32; 2002, n = 25; 2003, n = 26; 2004, n = 14. (A) Chi-square, 2001–2002 versus 2003–2004 = 34.0 (p < 0.001). (B) Mean ranks: 2001–2002 versus 2003–2004, 60.5 versus 29.3, p < 0.001, Kruskal–Wallis; 2001 versus 2002, p = 0.04; 2001 versus 2003, p < 0.001; 2001 versus 2004, p = 0.005; 2002 versus 2003, p < 0.001; 2002 versus 2004, p = 0.01, Mann–Whitney U-test.

Figure 2

Percentage of maternal and umbilical cord blood levels with chlorpyrifos concentrations > LOD (A) and mean chlorpyrifos ranks (B) by year of sample collection. Maternal blood sample size: 2001, n = 29; 2002, n = 22; 2003, n = 26; 2004, n = 15. Cord blood sample size: 2001, n = 21; 2002, n = 14; 2003, n = 20; 2004, n = 10. (A) In maternal blood, 2001–2002 versus 2003–2004, χ² = 12.2 (p < 0.001); in blood, 2001–2002 versus 2003–2004, χ² = 7.8 (p = 0.005). (B) Mean ranks, maternal blood: 2001–2002 versus 2003–2004, 51.7 versus 40.0, p = 0.001, Kruskal–Wallis; 2001 versus 2003, p = 0.008; 2001 versus 2004, p = 0.04; 2002 versus 2003, p = 0.005, Mann–Whitney U-test. Mean ranks, cord blood, 2001–2002 versus 2003–2004, 36.4 versus 29.0, p = 0.006, Kruskal–Wallis; 2001 versus 2003, p = 0.04, Mann–Whitney U-test.

Figure 3

Percentage of meconium samples with TCPy concentrations > LOD (A) and mean TCPy ranks (B) by year of sample collection. Sample size: 2001, n = 28; 2002, n = 21; 2003, n = 21; 2004, n = 13. (A) For samples > LOD, 2001–2002 versus 2003–2004, χ² = 22.1, p < 0.001. (B) Mean ranks, 2001–2002 versus 2003–2004, 50.0 versus 30.5, p < 0.001, Kruskal–Wallis; 2001 versus 2003, p < 0.001; 2001 versus 2004, p = 0.001; 2002 versus 2003, p = 0.02, Mann–Whitney U-test.