Preliminary epidemiologic investigation of the relationship between the presence of ambient hazardous air pollutants (HAPs) and cancer incidence in Harris County.

About the study…

This study was initiated by the City of Houston Health Department and conducted by the University of Texas School of Public Health. 

This study is a preliminary epidemiologic investigation of the relationship between the presence of ambient hazardous air pollutants (HAPs) and cancer incidence in Harris County. The cancer investigated is lymphohematopoietic.

The investigation consists of two analyses:

In the first analysis, home address in proximity to the Houston Ship Channel (HSC) is used as a proxy for exposure to HAPs.  The rationale for this proxy is the large presence of major industrial HAP emission sources along the ship channel. The difference between cancer incidence rates in people living near the ship channel is compared with those living farther away from the ship channel.

In the second analysis, the average ambient air concentration of the individual hazardous air pollutants benzene and 1,3 butadiene are used as a proxy for exposure to the HAPs family.  The rationale for this proxy is that these specific HAPs are two of the top seven air toxics identified, in a probabilistic human health risk screening study released by the Mayor’s Task Force in the spring of 2006, as posing a definite risk to human health in Houston.  These chemicals are found in high concentration in the HSC area. 

In general, the main objective of an epidemiological study is to find out if, within a group of people, or cohort, there is evidence of a cause and effect relationship between the incidence of disease and exposure to a chemical or physical agent.  In the study the health or disease patterns in a population group exposed to the chemical or agent, are compared with those of an unexposed or lesser exposed “control” group.  The differences in the numbers of the sick vs. healthy people between the groups are statistically analyzed to determine if the differences between the number of sick people within the groups are those that would likely occur by chance alone or if in fact, the disease pattern differences occur outside the realm of those expected from chance and are caused by exposure to the chemical.

What this study does do?

1. The strongest and most important findings of this study are the trend of increased incidence of:

• a particular cancer (acute lymphocytic leukemia) in children living in Harris County in close proximity to the Houston Ship Channel compared to those living farther away and
• any type of leukemia in children living in Harris County in regions with higher average ambient air 1,3 butadiene concentrations compared to those with lower average ambient air 1,3 butadiene concentrations.

2. The link between leukemia and 1,3 butadiene exposure association found in this study is consistent with and supports the findings of many previous studies in other geographic regions.  It is the first study of this type to relate incidence of cancer and these HAPs in Harris County.

3. This study underscores and supports the need for control of HAPs in ambient air.

4. This study emphasizes the need for a more in depth epidemiological study which better handles impacts of meteorology, refined source location, ambient concentration statistics and identified confounding factors.

What this study does not do?

1. This study does not provide the first or new insight into the carcinogenicity of these two HAPs.  EPA has previously characterized benzene and 1, 3 butadiene as human carcinogens based on the weight of evidence from many sources.

2. Because of the complexity of this type of study and the many levels of inherent uncertainty, it does not and cannot identify a bright line of exact concentrations or geographic boundaries associated with increased cancer incidence.

3. Although the study did not find consistent statistically significant overall trends in increased cancer incidence associated with adults:

• living in close proximity to the ship channel compared with those living farther from the ship channel and
• living in Harris County in regions with higher average ambient 1,3 butadiene and benzene concentrations compared to those with lower average ambient air 1,3 butadiene and benzene concentrations,

the study results cannot be extrapolated to indicate that there is no such relationship.  A more detailed assessment of the statistical power of the analysis would need to be conducted to ascertain if there was enough data to reject the hypothesis of increased cancer with increased exposure to HAPs.  In addition, it is possible that the relationship does exist but cannot be seen with this preliminary assessment.  There are many confounding factors which may not be completely controlled with this assessment.  A confounding factor is a particular circumstance that could influence the outcome of the study.  The statistics must include control or correction for the impact of confounding factors so that there is a clear path between the chemical and response. 

Association in adults is more complicated because of additional confounding factors more easily accounted for in children.  For example, mobility or changing residences throughout life ( the study assumes that the address when cancer was reported is the lifetime address although there is a long latency period of cancer), occupational exposure (comparing male and female rates may tease out impacts from occupational exposure however this requires more observations to attain statistical power), lifestyle choices (e.g., smoking) may not be accurately corrected, the ambient concentrations used in this study may not be representative of the actual exposure concentration/duration causing the cancer because of  long  and varying latency periods.

4. Although the study did not find consistent statistically significant overall trends in increased cancer incidence associated with children living in Harris County in regions with higher average ambient benzene concentrations compared to those with lower average ambient air benzene concentrations, the study results cannot be extrapolated to indicate that there is no such relationship.  In addition to statistical power issues and adequate control of confounding factors as discussed above, the differences in results found for 1,3 butadiene compared to benzene may stem from other causes such as the calculation of the exposure statistic and the nature of the chemical (e.g., 1,3 butadiene is a highly reactive chemical that is localized around its area of production in the HSC region while benzene is a ubiquitous pollutant in Harris County and a more refined concentration estimation procedure incorporating variability and confidence is recommended).

What is the City of Houston currently doing about this problem?

In addition to the routine oversight and enforcement of air pollution sources conducted by the City's Bureau of Air Quality Control, over the past 12 to 18 months, the City of Houston has initiated specific actions to address 1,3-butadiene and benzene air pollution problems through many avenues including:

• Negotiations of emission reductions with specific facilities emitting 1,3-butadiene: The City entered into an emission reduction agreement with the facility that emits the highest quantity of 1,3-butadiene emissions in the City so that reductions of emissions are quantified and verified;.
• Air monitoring and data surveillance of these pollutants has been increased: The Milby Park canister monitor was upgraded to an automatic gas chromatograph so that near real time hourly data are available. The data are routinely reviewed for spikes and trends.  A mobile air pollution monitor and laboratory is slated to be online in May to quantify impacts from specific 1,3-butadiene and benzene sources.
• The City has advocated for legislative and regulatory improvements to include: state ambient air pollutant standards for these and other air toxics; more stringent source emission controls from EPA.  
• The City is implementing a major source voluntary emissions reduction plan for benzene.