Concurrent Fetal Exposure to Multiple Environmental Chemicals along the U.S.—Mexico Border: An Exploratory Study in Brownsville, Texas

There is mounting concern that cumulative exposure to diverse chemicals in the environment may contribute to observed adverse health outcomes in the Lower Rio Grande Valley of Texas. To investigate this situation, biomarker concentrations of organochlorine (OC) pesticides/metabolites, polychlorinated biphenyls (PCBs), and polycyclic aromatic hydrocarbons (PAHs) were measured in maternal and umbilical cord blood from pregnant Hispanic women in Brownsville, TX. Results show that both mothers and fetuses were exposed concurrently to a variety of relatively low-level, hazardous environmental chemicals. Approximately 10% of the blood specimens had comparatively high concentrations of specific OC pesticides, PCBs and PAHs. Because many pregnant women in Brownsville live in socioeconomically-disadvantaged and environmentally-challenging circumstances, there is appropriate concern that exposure to these exogenous substances, either individually or in combination, may contribute to endemic health problems in this population, including cardiovascular disease, obesity, and diabetes. The challenge is to identify individuals at highest comparative risk and then implement effective programs to either prevent or reduce cumulative exposures that pose significant health-related threats.

One region where prenatal exposure to multiple environmental chemicals is a particular concern is along the U.S.-Mexico Border in the Lower Rio Grande Valley (LRGV) of Texas. Although actual exposure measurements are scarce, there is strong presumptive evidence that maternal/fetal exposures occur because of inadequate water supplies, marginal or nonexistent sewer systems, both indoor and outdoor air pollution, pesticide use, inappropriate waste disposal, unsafe food, and household and workplace use of hazardous chemicals [35][36][37][38][39][40]. Adding to the substantial pollution burden are a plethora of nonchemical stressors, including: rapid growth of a majority Hispanic, primarily Mexican American population; low rates of educational attainment; high rates of unemployment and poverty; a shortage of health care providers; and a general lack of awareness and knowledge about environmental health issues. Among the environmentally-related diseases and health outcomes that tend to be elevated in the LRGV are chronic liver disease, hepatitis A, tuberculosis, obesity, metabolic syndrome, diabetes, and hypertension [41][42][43][44][45][46].
In [2005][2006], an exploratory study to investigate the magnitude of concurrent fetal exposures to OC pesticides, PCBs, and PAHs (See Table 1 for common exposure routes) was undertaken in Brownsville, Texas; a city in the LRGV at the eastern end of the U.S.-Mexico border. According to the U.S. Census Bureau [47], Brownsville has a population of 172,434, with 92% of residents classified as Hispanic.
In 2006, it ranked as the most impoverished city in the U.S. based on average-annual household income, and 45% of residents who were 18-years-old or younger lived in poverty-the highest proportion of any city in the country with a population over 100,000 [47]. In this article, we report on results from a biomarker study of 35 pregnant Hispanic women residing in Brownsville, including postpartum analysis of cord blood, and summarize contemporaneous biomarker measurements of 30 OC pesticides/metabolites, 22 PCB compounds or groups, and 55 PAHs.

Subjects and Methods
The study participants were recruited from pregnant women in their first or second trimester presenting at a private gynecological clinic in Brownsville. Informed verbal and written consent, either in English or Spanish, as appropriate, was obtained, and no incentives were provided to research subjects. At the time of enrollment participants completed a brief questionnaire on demographic and socioeconomic characteristics. The study was approved by the Committee for the Protection of Human Subjects at the University of Texas Health Science Center at Houston. Previous publications describe results specific for PAHs [18], OC pesticides [7], and PCBS [13].

Study Participants
The sociodemographic attributes of women participating in the study are summarized in Table 2. All 35 women were Hispanic, between the ages of 18 and 38 years old, and had lived in the Brownsville area for several years. Only one woman identified herself as an active smoker, but her blood PAH concentrations were unremarkable. Average height was 5.2 ft, average weight was 161 lbs, and the women, on average, had 2.6 children at home and had experienced 2.8 previous pregnancies. Sixty-three percent were born in U.S., 6% in Mexico, and 31% elsewhere or unknown. More than a quarter (28.6%) did not finish high school, and 48.6% graduated from a college or university. Sixty-three percent worked outside the home and 20% described themselves as "housewives". Most women (71.4%) were married, while 25.7% had never been married and 2.9% were separated.

Specimen Collection and Handling
Sample collection occurred between October 2005 and February 2006, with venous blood samples collected during routine third-semester clinical visits and cord blood obtained at birth. Variable time elapsed between collection of maternal blood and cord blood, with 6 matched maternal-cord sample pairs obtained within 24 h of each other, 10 within 2-14 days, 16 within 15-35 days, and 3 within 43-57 days. Venipuncture was used to take maternal blood and samples were put into 10 mL, red-topped, vacutainer tubes, then labeled and refrigerated. The umbilical cord was severed at birth and approximately 10 mL of cord blood was drained into a red-topped, plain, vacutainer tube, which was capped, labeled, and refrigerated. Each unopened blood tube was sealed with Teflon tape and placed upright in an individual slot inside a pressure jar for shipping. Gel or ice packs were placed under, around, and over the jar, which was then sealed in a shipping container and sent by overnight express to the laboratory.

Analytical Methods
The laboratory at the Texas A&M School of Rural Public Health in College Station, TX, analyzed all samples [48]. Blood samples were analyzed (a) for PAHs using a combination of gas chromatography and mass spectrometry (GC/MS) according to modified U.S. Environmental Protection Agency (EPA) SW-846 Method 8271C [49], and (b) for OC pesticides and PCBs using gas chromatography with an electron capture detector (GC/ECD) according to modified U.S. EPA SW-846 Method 8081A [49]. All results are reported on a wet-weight basis (i.e., per unit of serum). Details about sample preparation, limits of detection, procedures for dealing with not-detected and below-detection-limit values, and reporting of wet-weight versus lipid-adjusted concentrations have been published separately for PAHs [18], OC pesticides [7], and PCBs [13].

Results and Discussion
Maternal and cord blood specimens were analyzed for 55 PAHs, 30 OC pesticides and/or metabolites, and 22 PCBs. As shown in Table 3, 19 PAHs (34.5%) were not detected at all and 10 PAHs (18%) were not detected in ≥75% of samples. Similarly for OC pesticides and PCBs, 14 OC pesticides/metabolites (47%) and six PCBs (27%) were not detected in any specimens and 11 OC pesticides/metabolites (37%) and nine PCBs (41%) were not detected in ≥75% of samples. These compounds (39 PAHs, 25 OC pesticides/metabolites, 15 PCBs) were excluded from further statistical analysis, but are included, when appropriate, as part of "total PAH", "total DDT", and "total PCB" values reported in Tables 4-6.
Geometric means and standard deviations for all compounds measured in >25% of maternal and cord blood specimens (26 PAHs, 5 OC pesticides/metabolites, 7 PCBs) are summarized in Table 4, along with the ratio of cord-to-maternal blood concentrations. Observed differences between cord and maternal blood were statistically significant (p < 0.05) based on two-tailed t-tests for 17 PAHs and for total PAH. The nine PAHs for which concentrations were not statistically different were naphthalene, C1-naphthalenes, biphenyl, acenaphthene, C3-fluorenes, 1-methylnaphthalene, 2-methylnaphthalene, 2,6-dimethylnaphthalene, and 1,6,7-trimethylnaphthalene. Differences between cord and maternal blood were not statistically significant (p < 0.05) for any OC pesticides/metabolites and PCBs, including total HCH, total Chlordane, total DDT, and total PCB, based on calculated z-scores.
It is worth noting that, because 29 out of 35 cord-maternal blood pairs were collected days to weeks apart, a direct comparison of the cord-to-maternal PAH ratio implicitly assumes that PAH exposures were relatively constant over the time period that data were collected. There is no way to know for sure, but the same general pattern (i.e., cord-blood greater than maternal-blood concentrations) was also observed for all six matched-sample pairs collected less than 24 h apart.
Correlations between total PAHs, total DDT, and total PCBs are presented in Table 5. Based on Pearson correlation coefficients, correlations between total PAHs and total DDT and between total PAHs and total PCBs were not statistically significant (p < 0.05). In contrast, correlations between total DDT and total PCBs were positive and statistically significant (p < 0.05) in cord blood (n = 35), maternal blood (n = 35), and all cord and maternal specimens combined (n = 70). Table 3. Summary of analytes not included in the statistical analysis because they were either (a) not found at all or (b) not measurable in ≥ 75% of maternal and cord blood specimens.     Data summarized in Table 6 provide a comparison of selected analytes for the three highest (approximately the upper 10th percentile) measured maternal blood concentrations from Brownsville and a statistical sample of Mexican American adults as reported in the CDC's (Centers for Disease Control and Prevention's) Fourth National Report on Human Exposure to Environmental Chemicals [50]. Geometric mean values for OC pesticides and PCBs in the Brownsville sample were either comparable (heptachlor-expoxide, 4,4'-DDT, PCB 66, 90/101, 138, 132/153/168, 180) or lower (trans-nonachlor, 4,4'-DDE) than those reported by CDC. However, high-end concentrations in the Brownsville sample were higher than CDC means for heptachlor-expoxide, trans-nonachlor, 4,4'-DDE, 4,4'-DDT, and PCB 66, 90/101, 138, 132/153/168 and 180, and higher than CDC 95th percentile values for heptachlor-expoxide, trans-nonachlor, 4,4'-DDT, and PCB 66. The CDC report did not include blood concentrations of PAHs, and did not report values for total DDT or total PCBs.

Discussion
Pregnant women and their fetuses in Brownsville, TX are exposed routinely to a diverse mixture of PAHs, OC pesticides, and PCBs. Maternal biomarker concentrations are comparable to cord blood for OC pesticides and PCBS, while cord blood levels of PAHs tend to be slightly higher. The majority of individual PAHs, OC pesticides, and PCBs were either (a) not detected at all or (b) not detected in >75% of the specimens, and mean concentrations of measurable compounds were generally relatively low. Nonetheless, approximately the upper 10th percentile of the Brownsville sample had relatively high concentrations of certain analytes, including heptachlorpexpoxide, trans-nonachlor, 4,4'-DDE, 4,4'-DDT, PBC 66, PCB 90/101, PCB 138, PCB 132/153/168, and PCB 180. Measured concentrations of total PAHs and total PCBs were 5-to 10-fold higher than the highest individual compound for each specimen, while total DDT levels were comparable to measured 4,4'-DDE concentrations. Total PAH values were not correlated (p < 0.05) with either total PCBs or total DDT, but total DDT was correlated (p < 0.01) with total PCBs in both maternal and cord blood.
As always, results must be interpreted in the context of study design limitations. As this was an exploratory study, we used a comparatively small (n = 35) convenience sample of pregnant Hispanic women presenting at a private OBGYN clinic, which biased the sample toward the upper socioeconomic strata (based on education and occupation). However, while the sample may be more affluent than average for Brownsville, they still live and work in a general socioeconomic environment that is disadvantaged economically, socially, and politically. It is also not known how, if at all, results were affected by the time elapsed between the collection of maternal and cord blood specimens. Furthermore, relatively low concentrations of blood biomarkers in most blood samples meant that analytical instruments were operating at or below nominal limits of detection for many compounds.
This study is one of the first to measure exposure biomarkers in residents of Brownsville (pregnant women in this instance). More than 90% of the population in Brownsville is Hispanic, and many individuals and families have socioeconomic attributes associated with poor health outcomes, including poverty, illiteracy, English as a second language, inadequate housing, substandard diet, lack of access to health care, and overall more stressful and less healthful lives [41][42][43][44][45][46]. Many impoverished Hispanics living along the U.S.-Mexico border tend to lack adequate sewage facilities, experience frequent flooding, live without air conditioning, not have sidewalks and street lights, and have insufficient knowledge about healthy lifestyles. Among common environmental hazards encountered by this population are (a) contaminated drinking and recreational water from agricultural runoff, municipal waste, and factory discharges, (b) adulterated soil from pesticide use, waste disposal, and illicit scrap yards and tire dumps, (c) polluted air caused by emissions from motor vehicles, industrial plants, and agricultural operations, and (d) tainted food from unhygienic growing, handling, processing, and cooking, as well as from subsistence farming and fishing [35][36][37][38][39][40]. Brownsville has a higher-than-average prevalence of several environmentally-related chronic diseases, including cardiovascular disease, tuberculosis, obesity, and diabetes [41][42][43][44][45][46], and children living along the border are hospitalized with asthma at a 36% higher rate than non-border children [51].
PCBs and OC pesticides are produced by industrial manufacturing processes, and both are toxic to humans, persist in the environment for years, bioaccumulate in the human body, and have a tendency to biomagnify in the food chain [50,52]. Consequently, their use has been banned or restricted in most industrialized countries for many years. The 2001 Stockholm Convention [52], a global treaty signed by more than 150 countries, required the elimination of production and use of PCBs and 13 OC pesticides. It also restricted the manufacture and application of DDT. But the reality is that because of continued use by some countries, along with long-range environmental transport, OC pesticides and PCBs are still routinely measured in air, water, soil, sediment, fish, birds, and mammals from all over the world [50,52]. PAHs, on the other hand, are inadvertent by-products of incomplete combustion and are virtually ubiquitous in urban and suburban settings [50,53,54]. Although PAHs are relatively short-lived in the environment, pregnant women in the U.S. are regularly exposed to PAHs in air, water, food, beverages, dust and soil [15,25,50].

Conclusions
Results from this exploratory study have demonstrated that even in Brownsville's higher socioeconomic strata (based on education and occupation), approximately 10% of pregnant Hispanic women in our sample had elevated levels of certain PAHs, OC pesticides, and PCBs in both maternal and cord blood. It is possible, if not likely, that exposures are even higher for those who are less well-off and unable to afford prenatal care at a private clinic. Given the socioeconomic privation of the Brownsville population, and the higher prevalence of many chronic diseases, follow-up research should be conducted to (a) establish the distribution of exposures, with particular emphasis on the poorest and least educated residents, (b) identify sources and pathways for those at the high-end of the exposure distribution, and (c) explore associations between exposures and prevalent chronic conditions in this population, like obesity and diabetes. A suitably large, statistically-based sample should be used and more detailed information on participants' diets, housing characteristics, neighborhood settings, living circumstances, and activity patterns should be obtained. The overall goal should be to examine cumulative fetal exposure to multiple chemical and nonchemical stressors in the environment, with particular focus on identifying those individuals at highest comparative risk for adverse health outcomes.