Hazardous Air Pollutants Special Report

While recent reports have commented on the elevated ambient levels of hazardous air pollutants (HAPs) in certain areas of Houston, relative to other cities, few studies have assessed the health effects of HAPs for Houstonians and none have evaluated the association between ambient levels of these pollutants and lymphohematopoietic cancer risk in this population. To begin to address this deficit in the literature, we conducted a population based analysis of ambient levels of HAPs and lymphohematopoietic cancer incidence in Harris County. Cancer cases were obtained from the Texas Cancer Registry (TCR), 1995-2003, and included adult and childhood cancers. We used existing monitoring data collected from 1992-2003 by the Texas Commission on Environmental Quality (TCEQ) to estimate ambient benzene and 1,3 butadiene levels for each census tract, which were then collapsed into 3 or 4 categories, respectively, in the analyses that were carried out. We also estimated risk of lymphohematopoietic cancer associated with residential proximity to the Houston Ship Channel (HSC) as another proxy (albeit a crude measure) of HAP exposure. Poisson regression was used to estimate rate ratios for increasing HAP levels and cancer rates, adjusting for gender, age, and a composite index for socioeconomic status and race / ethnicity. We observed an association among children living within two miles of the HSC and higher rates of acute lymphocytic leukemia (adjusted rate ratio [aRR] = 1.56; p=0.01) compared with those living greater than 10 miles from the HSC. Further, higher 1,3-butadiene levels (>1.15 parts per billion [ppbV] relative to <0.266 ppbV) were associated with acute lymphocytic leukemia, acute myeloid leukemia and all leukemias in children. While we did not observe statistically significant associations for the other two categories (i.e., 0.266 0.381 ppbV or 0.382 1.15 pbbV) as compared to the referent group of <0.226 ppbV, a trend of increasing 1,3-butadiene levels and increasing leukemia rates was noted for these cancers. Higher benzene levels were not associated with lymphohematopoietic cancer by types in children. Among adults, neither proximity to the HSC, nor ambient levels of benzene or 1,3-butadiene levels were associated with lymphohematopoietic cancers in a dose dependent manner for either males or females. Additional analyses using more sophisticated methods to assess exposure are planned to confirm our findings. Additionally, future analyses will address many of the limitations of the current analysis. However, observing a specific health effect of HAPs in light of recently documented elevated levels of two known carcinogens, benzene and 1,3 butadiene, in Houston,(1) suggests a need to explore this issue further and to potentially take action to limit potential exposure to HAPs. Introduction In 1997, Texas ranked among the top three states in the number of chemical and allied products manufacturing facilities and in the number of petroleum and coal products manufacturing facilities, with more facilities located in Harris County (Houston) than in any other Texas county. (2) In a recent report on the effects of air pollution, the Mayor’s Task Force determined that ambient levels of both benzene and 1,3-butadiene pose a ‘definite’ cancer risk to Houstonians. (1) According to the International Agency for Research on Cancer (IARC), benzene is classified as a known human carcinogen and 1,3-butadiene is classified as a probable human carcinogen. (3) Few studies have assessed the health effects of hazardous air pollutants (HAP) for Houstonians and none have addressed ambient levels of these pollutants and risk of lymphohematopoietic cancer, which includes leukemia, lymphoma and myeloma. The majority of studies that have examined the potential health effects of HAPs have been conducted in occupational workers. However, most have not directly measured individual exposures to pollutants. The HAPs of primary concern in Houston, benzene and 1,3-butadiene, are mainly produced by petrochemical and synthetic rubber and plastics manufacturers. Occupational studies conducted in these settings typically compare worker populations with the population at large and report lower cancer mortality and incidence rates, for all cancers combined. (4-23) Four studies reported more deaths from lymphohematopoietic cancers than expected among refinery/petrochemical plant employees (10, 13) or those exposed to 1,3-butadiene (7) or benzene (24) while findings from eight studies do not support this association.(6, 16, 17, 19, 21, 25-27) When looking at risk of developing cancer, four worker based cohort studies found higher risk of leukemia associated with exposure to benzene, (28, 29) or 1,3-butadiene (5, 30) and ten studies did not find an association. (6, 8-10, 12, 13, 16, 19-21) A nested case-control study (31) reported a significant increased risk of leukemia associated with an average exposure to 1,3-butadiene at levels of 1 ppm and another study reported that workers exposed to benzene at levels below 1 ppm had significantly decreased levels of white blood cells compared with unexposed workers. (32) Few studies have looked at HAPs and risk of cancer in community populations. In these studies, exposure to HAPs is usually measured with a proxy measure such as residential distance from a plant or spill. One study compared actual and expected number of cancers in an area affected by a gasoline spill as a measure of high exposure; residents’ leukemia rates were 4.40 (95% CI: 1.09-10.24) higher than expected by chance. (33) Another study found no effect of proximity (three kilometer area radius) to a petrochemical plant on the risk for leukemia (standardized incidence ratio [SIR]=0.99, 95% CI: 0.66-1.51) yet they did note a significant increased risk of multiple myeloma (SIR=2.15, 95% CI: 1.253.67). (34) Three other studies found no increased risk of leukemia in areas thought to be affected by industrial pollution.(35-37) Five have found a significant association with residence near industrial facilities and specific lymphohematopoietic cancers (38-42) while three others have found no association (34, 42, 43) Only a small number of studies have explored the potential association between HAPs and childhood cancers. In a large California based study including 7,143 childhood cancer cases, Reynolds et al. (15) found a slight but non significant increase in leukemias (aRR=1.15; 95% CI 0.97–1.37) and proximity to road ways; there was no evidence that cancer rates increase with increasing vehicle or road density. In a follow-up for this study, Reynolds et al. (14) used data from the same 7,143 cases but used a dispersion model developed by the EPA to assign census tract level exposure scores to 25 potentially carcinogenic HAPs. Exposure scores for individual HAPs were summed to create a combined HAP exposure score. Reynolds et al. found a significant (p<0.05) trend of increasing childhood leukemia rates with increasing HAP exposure levels with a 21% (aRR = 1.21; 95% CI: 1.03-1.42) increased risk of childhood leukemia in census tracts with the highest combined exposure score. They further report a 32% increase in leukemia risk (aRR = 1.32; 95% CI: 1.11-1.57) associated with the highest levels of HAP exposure from point sources, compared with non-significant leukemia risks associated with the highest levels of HAP exposures from mobile (aRR=1.18) and area (aRR=1.16) sources. In the United Kingdom, Knox found correlations between residence at birth near geographically defined “hotspots” for several criteria pollutants as well as volatile organic compounds (VOCs), benzene, dioxins, 1,3butadiene, and benz(a)pyrene and childhood leukemias. (11) Additional studies are needed to confirm the findings from these studies. Because there have been few community level studies of HAP and cancer risk and because most of these studies have not utilized data on ambient levels of specific air pollutants, we have conducted a population based analysis of the association between HAPs and lymphohematopoietic cancer incidence in Harris County. We used existing monitoring data collected from 1992-2003 by the Texas Commission on Environmental Quality (TCEQ) to estimate census tract ambient levels of benzene and 1,3-butadiene. Cancer cases were obtained from the Texas Cancer Registry (TCR), 1995-2003, and included cases in adults and children.