• Home
• About ATSDR
• Press Room
• A-Z Index
• Glossary
• Employment
• Contact Us
• CDC

ATSDR en Español

Search:

Skip directly to: content | left navigation | search

---

Find a Publication

1. ATSDR > Public Health Assessments & Consultations

PUBLIC HEALTH ASSESSMENT

PANTEX PLANT
AMARILLO, CARSON COUNTY, TEXAS

IV. ENVIRONMENTAL CONTAMINATION AND OTHER HAZARDS

IV.A Introduction

The Agency for Toxic Substances and Disease Registry (ATSDR) selects and discusses contaminants based on several factors: sample design, field and laboratory data quality, and comparison of chemical concentrations to levels that could cause cancer or other health effects. In addition, community health concerns are considered.

Evaluating the sample design involved reviewing Department of Energy General Environmental Protection Program regulations as per DOE Order 5400.1, and the Pantex Plant's approach to locating and defining contamination. Spatial distribution of sampling locations, sampling frequency, concentration changes over time, medium-to-medium differences, and correlation between the selected list of analytical parameters and suspected environmental contaminants are factors considered by ATSDR when determining the contaminants to which humans could be exposed.

Review of field sampling quality control procedures included interpreting data on background (or regional) concentrations of chemicals. Additionally, the adequacy and number of replicate, spiked, and blank samples are checked to verify detection of contaminants. To assess laboratory quality control, procedures used to verify instrument reliability are reviewed.

Contaminant concentrations detected on- and off-site are compared to values that are believed to be without adverse health effects upon exposure. Those values are developed by health and regulatory agencies to provide estimates of levels at which health effects might be observed. Those values, in many cases, have been derived from animal studies. Health effects are related not only to the exposure dose, but to the route of entry into the body and the amount of chemical absorbed by the body. For those reasons, comparison values used in public health assessments are contaminant concentrations in specific media and for specific exposure routes. Several comparison values may be available for a specific contaminant. ATSDR generally selects the comparison value that is calculated using the most conservative exposure assumptions in order to protect the most sensitive segment of the population. The potential for adverse health effects from exposure to contaminants is discussed in the Public Health Implications Section of this document.

ATSDR reviewed the non-radiological (chemical) and radiological data collected in support of environmental monitoring at the Pantex Plant. The material reviewed covered the period of 1973 through 1998. A combined list of environmental monitoring parameters for 1992 through 1997 is provided in Appendix A. The documents reviewed included annual and supplemental environmental reports, support documentation for the environmental impact statement (EIS) prepared by the Los Alamos National Laboratory, and the on-going EIS.

IV.B Toxic Chemical Release Inventory

Under Section 313 of the Emergency Planning and Community Right-to-Know Act (SARA, Title III), manufacturers are required to report to the Environmental Protection Agency (EPA) annually if they have released into the environment (routinely or accidentally) any of more than 300 toxic chemicals. Section 313 authorizes EPA to maintain the data in a computerized database that is known as the Toxic Chemical Release Inventory (TRI). Manufacturing facilities (as defined in the Standard Industrial Classification codes 20-39) that have 10 or more full-time employees and that manufacture or use a Title III-listed chemical in an amount greater than its specified threshold for manufacture, import, processing, or other use during any calendar year are required to estimate their annual releases of such toxic chemicals into the air, water, and land. The database is available to federal and state government officials and to the public.

A review of the current TRI file database for the Pantex Plant, consisting of data for calendar years 1987-1994, did not indicate the occurrence of environmental releases expected to cause adverse health effects.

IV.C Environmental Contamination

IV.C.1 Terminology Definitions

The following terms are used in this section:

CMS (Corrective Measures Study) The analysis of potential cleanup methods and is part of the corrective action process under the Resource Conservation and Recovery Act (RCRA). The RCRA CMS is equivalent to the Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA) Feasibility Study.

EMEG [Environmental Media Evaluation Guide (ATSDR)] EMEGs are screening values used to select chemical contaminants for further evaluation in the ATSDR public health assessment. EMEG values are calculated by ATSDR using conservative exposure assumptions that would protect the most sensitive segment of the population.

ICM (Interim Corrective Measure) A remedial action employed in the interest of protecting human health and the environment to address the hazards posed by actual or suspected contamination.

MCL (Maximum Contaminant Level) MCLs represent contaminant concentrations that EPA deems protective of public health over a lifetime (70 years) at an exposure rate of two liters of water per day. MCLs are also regulatory concentrations.

mg/Kg [milligrams per kilogram (parts per million)] The unit applied to express contaminant concentrations in soil.

NFA (Finding of No Further Action) A recommendation of no further remediation action based on the finding of no contamination, or the finding that remediation of contaminants is not warranted based on Risk Reduction Standards as established by the Texas Natural Resource Conservation Commission (TNRCC).

n/a not available.

Phase I Fieldwork Activity performed to determine whether contamination exists and, if so, the vertical and lateral extent of the contamination.

Phase II Fieldwork Activity performed if Phase I fieldwork identifies a need for additional information to determine the proper characterization of the site.

Risk Reduction Standards Guidance established by the TNRCC to ensure that Resource Conservation and Recovery Act (RCRA) remedial actions are completed with utmost regard to present and future threats posed to human health and the environment. The Risk Reduction Standards also draw on guidance from the Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA).(6)

SWMU (Solid Waste Management Unit) Any unit from which hazardous constituents may migrate irrespective of whether the units were intended for the management of solid and/or hazardous wastes.

SWMU Grouping A discrete portion of a DOE installation identified as a group of release sites treated as a single project for remediation. SWMUs are grouped by contaminant commonality for the purpose of identification.

µg/L [micrograms per liter of water (parts per billion)] The unit applied to express contaminant concentrations in water.

UST (Underground Storage Tank) A stationary device, designed to contain an accumulation of liquid waste, solid waste, or product, and constructed primarily of non-earthen materials providing structural support. Such a tank is classified as a UST when at least 10% of the waste/product volume is below the surface.

IV.C.2 Overview of the Pantex Plant Remediation Process

The Resource Conservation and Recovery Act (RCRA) Facility Investigation (RFI) is designed to detect the presence of hazardous substances and to determine whether they may be contaminating the surrounding environment. RFIs are a condition of Pantex Plant's RCRA permit issued by the Texas Natural Resources Conservation Commission (TNRCC).

In January 1989, the EPA conducted a RCRA facility assessment to identify Solid Waste Management Units (SWMUs) that may require investigation and/or corrective action under the 1984 Hazardous and Solid Waste Amendments to the RCRA. The Draft Administrative Order of Consent, pursuant to Section 3008(h) of RCRA, was issued in September 1989. It identified 144 SWMUs. This order outlines requirements for performing interim measures, RFIs, corrective measure studies, and corrective measure implementation at the Pantex Plant SWMUs. The SWMUs were organized into 15 SWMU Groupings that are to be characterized and, if found to be contaminated, carried through the corrective action process on independent schedules. SWMU Activity Data Sheet (ADS) numbers are also provided.

The following is a list of the 15 Pantex Plant SWMU Groupings: (7)

A general description and comment on the current activity status of each of the Pantex Plant SWMUs is presented in Appendix D.

IV.C.3 Media-Specific Contamination (Non-Radiological) (3)

IV.C.3.a Groundwater

The Pantex Plant Groundwater Monitoring Program includes 77 wells (64 monitoring wells, 5 drinking water production wells, 7 water treatment study wells, and 1 off-site control location). Twelve monitoring wells, the single off-site control location, and all five production wells are completed in the Ogallala aquifer. The remaining 59 wells are completed in the perched aquifer.

Samples from the production wells are used as a secondary check for possible chemical migration from the perched aquifer into the Ogallala aquifer. If contamination of the Ogallala aquifer were to occur as a result of Pantex Plant operations, it would be discovered first in the production wells. In this manner, the Pantex Plant production wells serve as interceptor wells for the downgradient City of Amarillo drinking water production wells. The Pantex Plant and City of Amarillo production wells are also monitored in accordance with Safe Drinking Water Act requirements for source water.

The water supply at a control location in Bushland, Texas, is sampled monthly to obtain control data for comparison with data from the plant. This control well is completed in the Ogallala aquifer upgradient from Pantex Plant and is not affected by plant operations.

Constituents analyzed under the environmental monitoring programs for all media are listed in Appendix A. The analyses include four screening tests for indicators of contamination: pH, conductivity, total organic carbon, and total organic halogen. Analyses also include the water quality parameters specified under Resource Conservation and Recovery Act (RCRA) monitoring requirements: iron, manganese, sodium, chlorides, phenols, and sulfates.

Analytical results are evaluated quarterly and compared to guidelines for RCRA-regulated facilities, drinking water standards, and DOE criteria. Mean values between monitored wells and the off-site or control location are compared statistically to determine potential impacts of Plant operations on groundwater. Data are compared to the Safe Drinking Water Act Maximum Contaminant Levels (MCLs), Maximum Contaminant Level Goals (MCLGs), the risk reduction levels developed for Pantex Plant, or derived concentration guides for radionuclides. The risk reduction levels are generally at least as stringent as the Safe Drinking Water Act MCLs and also set standards for contaminants not regulated by the Safe Drinking Water Act.

Fifty-one on-site wells completed in the perched aquifer are sampled quarterly, semi-annually, or annually. Samples are analyzed for radiological and chemical constituents. No pesticides, herbicides, polychlorinated biphenyls (PCBs), or dioxins have been detected in perched groundwater. Chemical contamination initially detected in the on-site perched aquifer in the early 1990s and still present is made up of metals, explosive compounds, and volatile organic compounds. Because that particular perched aquifer is not utilized as a drinking water source downgradient of the contaminant plumes, no one is exposed to those contaminants. Pantex Plant is currently conducting perched aquifer remediation activities using pump, treat, and recharge technology to reduce contaminant concentrations to acceptable levels. A May 1998 update on the treatability study is provided in Appendix C.

In late 1995, a perched aquifer monitoring well in the southeastern corner of the Pantex Plant revealed high levels (in excess of 5,000 ug/L) of the explosive RDX (also known as Royal Demolition Explosive, Cyclonite, or Hexogen). Two new monitoring wells were installed in the perched aquifer on the off-site property (1,200 feet east of the Pantex Plant boundary) in May 1996. Sampling of those off-site wells in June 1996 and January 1997 detected explosive constituents at levels above ATSDR comparison values; however, it is critical to note that these are perched aquifer monitoring wells and this particular perched aquifer is not used as a drinking water source.

Two Ogallala wells were present at a property owner's home 1,200 feet east southeast of the two off-site monitoring wells. The first well (windmill well) was drilled to the Ogallala in 1896 as a drinking water well. In 1964, due to dropping Ogallala water levels, a new, deeper well (domestic well) was installed in the Ogallala. The new well was located 10 feet north of the old windmill well. The common problem with the two drinking water wells was in the manner of installation. The well casing ran from the surface to the Ogallala, where it was gravel packed for production. There was no barrier to inhibit flow from the perched zone (if it is or has been present and saturated) into the Ogallala through the annular space (the space between the casing and the bore hole). Apparently this was the case because, in May 1995 and again in May 1996, the Texas Natural Resource Conservation Commission (TNRCC) sampled the two drinking water wells and detected low concentrations of explosives at levels below health screening values.

To prevent further intrusion into the Ogallala, the two off-site drinking water wells were sealed and a new drinking water well was installed in the Ogallala in May 1996. Precautions were taken in the construction of the new well to prevent intrusion of contaminated perched aquifer water into the Ogallala. Particulate and carbon filters were also installed between the well and the property owner's home. Quarterly sampling of the new well has not detected contaminants at levels known to cause adverse health effects; however, contaminants such as metals are detected regularly and are typical of natural background levels.

Three monitoring wells in the Ogallala aquifer are sampled monthly and six quarterly for radiological and chemical constituents. Quarterly sampling in two additional wells was initiated in 1995. No chemical contaminants exceeding health screening values have been detected to date. Possible bacteriological contamination was detected in one of the production wells in July 1995; however, two subsequent samples did not detect and therefore did not confirm the presence of biological contamination.

TNRCC collected samples from the Pantex Plant drinking water supply system in October 1996 to determine whether regulatory requirements were met. Plant personnel collected additional samples at the same time. TNRCC also analyzed for lead and copper in samples collected in July 1996 by Pantex Plant personnel. Results for all analyses were similar and within regulatory limits.

In summary, although chemical contamination above screening values has been detected in the perched aquifer at the southeastern plant boundary, that particular perched aquifer is not used as a water source downgradient of the contaminant plumes. If that perched aquifer was used in the future as a water source downgradient of the contaminant plumes, water from the perched aquifer would require treatment to remove contamination before it would be safe for consumption. Chemical contamination above screening values has not been detected in the Ogallala aquifer, which is used as a drinking water source throughout the region.

IV.C.3.b Surface Waters

There are no rivers or streams at the Pantex Plant; the only surface waters occur in ditches and playas. Department of Energy (DOE) owned and DOE-leased property is associated with six playas, two of which are on the Texas Tech University DOE-leased property. Most stormwater runoff from the DOE-owned and DOE-leased lands flows to the on-site playas: 11 effluents from Plant operations are treated and, along with some noncontact industrial discharges, are directed into ditches that drain to three playas (Playas #1, #2, and #4), treated industrial discharges from Zone 12 and the eastern portion of Zone 11 flow through ditches to Playa #1, stormwater runoff from western portions of Zone 11 is channeled to Playa #2, stormwater runoff from the Burning Ground drains into Playa #3 as sheet flow, and runoff from southern portions of Zones 11 and 12 flows into Playa #4. Plant activities no longer influence flows to Pantex Lake (owned by DOE), located to the northeast because plant discharges to Pantex Lake were discontinued several years ago; however, monitoring is still conducted because of the past wastewater discharge practices. The area surrounding Pantex Lake is used for agriculture, and currently stormwater is the only contributor to water in the playa. Playa #5, which is on Texas Tech University property, is not influenced by operations at Pantex Plant and receives runoff from surrounding Texas Tech farming operations.

Surface water monitoring at the Pantex Plant is generally dependent on rainfall or discharge events; samples are collected when flows occur. Because flows are continuous at the Wastewater Treatment Facility, sampling there can be conducted on a regular schedule. Sampling at the Wastewater Treatment Facility was conducted in accordance with the Wastewater No-Discharge Permit issued in 1988 by the Texas Water Commission (the predecessor agency to the Texas Natural Resource Conservation Commission) until that permit was replaced with a Wastewater Discharge Permit issued by TNRCC June 14, 1996. A similar permit was issued to Pantex by EPA on June 1, 1996. For control purposes, off-site surface water is sampled from a playa at Bushland, Texas. The control data are compared with on-site data to identify significant differences in contaminant levels potentially resulting from plant operations. As shown in Appendix A, samples from selected monitoring locations are analyzed for metals, explosives, polychlorinated biphenyls (PCBs), pesticides, herbicides, volatile organic compounds, semivolatile organic compounds, and water quality indicators reflecting water quality and treatment.

The tail-water pit located in the northeast corner of Pantex Plant near the Old Sewage Treatment Plant (OSTP) receives runoff from the surrounding agricultural lands, the OSTP (inactive), and portions of the Firing Sites. The tail-water pit is a stormwater outfall, thus it receives no direct effluent discharge from the plant. Generally the tail-water pit contains all drainage from the northeast section of the plant and returns the water to Playa 1 by forced pumping; however, during high rainfall events, the pit can overflow into the drainage ditches along FM 293 which then channels the water to Pratt Lake, a playa located north of the Pantex Plant boundary. This is the only surface water runoff boundary area at Pantex Plant which may be influenced by stormwater runoff from operational areas.

The tail-water pit is currently sampled quarterly in addition to "special request" sampling events when the water overflows. The "special request" sampling is performed as a good-neighbor policy when plant personnel are present during an overflow event. Due to pumps located in the tail-water pit which return the water to Playa 1, overflow of this area occurs infrequently. The last 3 known overflow events occurred on February 14, 1997, April 25, 1997, and March 19, 1998.

The sampling program for the tail-water area has changed over time. Starting October 1997, the pit was (and is) sampled quarterly for visual contamination. Visual inspection parameters include color, odor, clarity, sheen, suspended solids, floating solids, settled solids and other obvious indicators of stormwater pollutants. Prior to October 1997, the tail-water pit was sampled semiannually and analyzed for water quality indicators, metals, volatile organic compounds, semi-volatile organic compounds, pesticides/herbicides, PCBs, chemical oxygen demand, pH, high explosives, radionuclides, and flow. The only parameter concentration detected above a background reference playa water sample was an occasional low-level detection of pesticide which would be reasonably expected from the surrounding agricultural areas. Due to a change in the NPDES permit to a new "Multi-Sector Storm Water Permit," sampling after October 1998 will be conducted quarterly and will include both visual and analytical sampling. The analytical parameters include total organic carbon, iron, gross alpha, and gross beta. For "special request" sampling, analytes include pesticides/herbicides, PCBs, volatile organic compounds, semi-volatile organic compounds, high explosives, metals, radionuclides, and water quality indicators.

Since the issuance of the Pantex permits, some metals have been detected at Outfall 006 at concentrations above permit guidelines. Those elevated concentrations resulted from the natural levels of metals in soil in the open ditches leading to the outfall weir. In a November 26, 1996, letter from J.S. Johnson (Pantex) to Ted Palit (EPA), "Written Notification of Non-Compliance with U.S. Environmental Protection Agency (EPA) Industrial Discharge National Pollutant Discharge Elimination System (NPDES) Permit No. TX0107107", the process for the identification, development, and implementation of corrective measures for elevated metal concentrations at this outfall was defined. Corrective measures identified and implemented included placement of gravel in localized areas affected by erosion upstream of Outfall 006, removal of all sediment from the drainage structure immediately upstream from the outfall with the exception of a thin layer that remains on the bottom of the channel and within sections of the security perimeter fences, and grading and placement of gravel in the buffer area immediately adjacent to the outfall structure.

Chlorine residual, ammonia, and total suspended solids have exceeded permit guidelines at the waste water treatment facility (WWTF); those elevated concentrations have been associated with algae growth in the lagoon. The concentrations of nitrogen, organic carbon, and phosphorus in the wastewater are high enough to promote significant algae growth, which is typical at a facultative lagoon. These nutrients, coupled with the seasonal increase in temperature and sunlight, enhance the growth of algae in the WWTF during the warmer seasons. Subsequently, algae growth is reduced by colder temperatures and decreased sunlight typical in the winter months.

No enforcement action has been issued by the TNRCC or EPA to date, but a compliance order to enforce existing permit limits was issued to the Pantex Plant by EPA. The EPA Administrative Order, which became effective December 1, 1997 requires negotiation of an action plan and schedule for correcting the violations. The violations mentioned in the Administrative Order included ammonia; oil, greasy, and total suspended solids; total metals (manganese, titanium, zinc, copper, and chromium from soil transported in storm water runoff); and silver from a photo processing operation.

The EPA and the Texas Natural Resource Conservation Commission inspected Pantex in April 1997 and identified no significant issues, with the exception of the permit violations. As a result of that inspection, Pantex has undertaken corrective actions to achieve compliance with the permit limits. Some of these actions include the algae-eating fish research project, initiated in October 1997, focused on controlling total suspended solid levels in discharge from the waste water treatment facility (WWTF); erosion and sediment control; and process modifications including the installation of a flow-weighted chlorine-dosing system. To further meet permit requirements, design of an upgrade to the existing WWTF was initiated in 1998. The goal of the project is to eliminate industrial discharges to open ditches and improve treatment capabilities.

The EPA and TNRCC have found no instances to date where contaminant concentrations have exceeded guidelines that pose a threat to human health. ATSDR did not identify exposure pathways related to the violations.

IV.C.3.c Soils

Routine surface soil surveillance at Pantex Plant is conducted both on-site and off-site with samples analyzed for metals, explosives, and volatile compounds (Appendix A). The routine soil surveillance program provides a direct measure of environmental contamination because soil accumulates contaminants deposited from the air and surface water over time; therefore, soil surveillance allows evaluation of long-term trends.

Soil is routinely sampled at 30 on-site and 17 off-site locations. In addition, samples are collected at 60 locations within Zone 4 West. All soil samples are collected according to sampling procedures outlined in Pantex Plant Internal Operating Procedure D4235, "Soil Sampling Procedures." Samples for radiological analyses are collected as 3.5 inch diameter by 2-inch deep cores. Samples for metals, explosives, and volatile organics are collected as 2-inch diameter by 4-inch deep cores.

Soil samples are collected from two general landscape positions: playa bottoms and interplaya uplands. This routine sampling of the playa bottoms began March 17, 1994. The characteristic soil types for these landscape positions are Randall clay in playa bottoms and Pullman clay loam in the interplaya uplands. Two off-site sampling locations (an upland location and a playa bottom) are located at Bushland, Texas. Samples collected at these two locations are considered controls, indicating background levels of contamination. The control locations were chosen because they are 35 miles from the Plant in a direction that is generally upwind and thus should not be affected by Plant operations. On-site soil samples are collected monthly from five areas: the Burning Ground, firing sites, Playa #1, Playa #2, and Playa #3. Samples are collected in Zone 4 West once during the year. Off-site samples are collected quarterly from areas within a 5-mile radius of the Plant. Samples from the control locations are collected monthly.

Because open burning has the potential to release metals, explosives, or volatile organic compounds that may be deposited or incorporated in surface water runoff, soil samples from the Burning Ground, Playa #3, and Bushland (control) are also analyzed for those contaminants. All samples are analyzed by an off-site contract laboratory that meets U.S. Environmental Protection Agency (EPA) requirements.

Overall, soil monitoring results have been within the observed ranges of concentrations for uncontaminated soil and comparable to historical results and those for the control locations. Although, boron, manganese, zinc, HMX, and total xylene concentrations in the area of the Burning Grounds exceeded screening values, the location is on-site in a controlled-access area thereby preventing exposure to unprotected individuals. The site is also being addressed under the site remediation program. Those compounds have not been detected off-site at levels that would cause adverse health effects.

IV.C.3.d Air

Pantex established a system of radiological air monitoring stations consisting of 9 continuous air monitoring stations at a 5 mile radius in November 1973. In 1995, concrete pads and electrical utilities at three of the fenceline air monitoring stations were augmented so that the stations could be used for both nonradiological and radiological monitoring. The three stations were upwind and downwind (based on the prevailing wind direction), respectively, of the Burning Grounds and were equipped to sample for hydrogen fluoride and volatile organic compounds. One downwind station was also equipped to sample for particulate matter less than 10 microns in size. The three sites provided an opportunity to compare Pantex Plant-collected data with data collected by the Texas Natural Resource Conservation Commission (TNRCC) on site.

Prior to 1998 TNRCC conducted routine air monitoring for air pollutants, such as volatile organic compounds, particulate matter less than 10 microns in diameter (PM-10), and hydrogen fluoride. Monitoring at those stations was conducted for 24 hours every sixth day on a scheduled rotation. In addition, a Fourier Transform Infrared monitor was installed in 1995 at the Masterson Pump Station, approximately 2 miles north by northeast of the plant. This station operated continuously, collecting data on volatile organic compounds and meteorological conditions.

TNRCC also conducted routine nonradiological ambient air monitoring at Pantex Plant, and results of the sampling program have been published in a series of quarterly reports. Except for a few days on site when particulate matter was detected at levels exceeding the National Ambient Air Quality Standard, and on one occasion when methylene chloride was detected at levels seven times the 24-hour Effects Screening Level, also on site, levels of pollutants detected have been below health screening values. TNRCC Effects Screening Levels (ESLs) are guideline comparison levels set below levels at which adverse health effects are reported in the scientific literature. If an air concentration of a chemical is below an ESL, adverse health effects are not expected to occur. However, if an air concentration is above the ESL, it is not indicative that adverse effects will occur, but rather that further evaluation is warranted. Particulate matter is occasionally present in air around Pantex Plant as a result of high winds and agricultural activity, and the maximum 24-hour methylene chloride concentration, although in excess of the screening value, was less than levels that would result in adverse health effects. The Air Quality Control Region including Pantex Plant and the surrounding Amarillo Area is classified by the Environmental Protection Agency (EPA) as "Attainment" for all National Ambient Air Quality Standards (40 CFR 52.2279). This classification means that the area meets all existing air quality standards for which it could be evaluated.

In 1997, budget constraints combined with monitoring data indicating no significant levels of air contamination from past air sampling, resulted in a decision to reduce the level of monitoring frequency beginning in 1998. It was determined that monitoring would be on an episodic, event-triggered basis, with a maximum of thirty sampling events per year at two sites (Sites No. 4 and 5) containing noncontinuous monitors. Continuous monitoring would continue as in the past at one site containing a continuous Fourier transform infrared (FTIR) monitor (Site No. 7). The TNRCC Regional Manager, located in the Amarillo Office, will determine the particular sampling day, based on criteria such as burning days, grass fires, or other work activities with potential impact on air quality. To assist in that effort, the Pantex Plant is required to provide a weekly schedule to TNRCC of any scheduled activity that could result in a potential impact on air quality as well as immediate notification of grass fires, upset or emergency maintenance situations, or any unscheduled activities potentially impacting air quality.

Parameters to be measured under the 1998 program include TSP, PM-10 and VOCs, using both gas chromatograph/mass spectrometer detector (GC/MSD) and FTIR detection. Meteorological parameters of wind speed, wind direction, temperature, and relative humidity will also be measured.

Monitoring sites are located around the burning ground of the Pantex facility to establish the effect of normal operating activities on the ambient air. Meteorological information obtained from the National Weather Service was used to establish the most prevalent wind direction, which is from the south/southwest. Based on this information, sampling sites were established around the burning grounds, and Zones 11 and 12, where most of the solvents are handled. An additional monitoring site, Site No. 7, is located to the northeast of the Pantex facility. Under the 1998 program, three of the five original sites remain operational, Sites No. 4, 5, and 7. At this time, there are no plans to dismantle the inactive sites.

Because all the air monitoring sites require 240 volts alternating current power, they are located near power lines and transformer locations. The TNRCC Ambient Air Monitoring staff coordinated with the DOE contractor to acquire needed utility connections and security provisions. Because of the sensitive nature of nuclear weapons facilities, security considerations of the DOE were also a determining factor in monitor locations.

Siting criteria (established by the U.S. Environmental Protection Agency in 40 Code of Federal Regulations Part 58, Appendices D and E) for ambient monitoring sites were followed. Therefore, monitoring sites are located where they are not hampered by tall structures or trees. Consideration was also given not to place TSP and PM-10 monitors close to unpaved roads or alleyways. Accessibility to the monitoring sites is another important consideration because each site is visited frequently to change filters or sampling canisters. Therefore, monitors were sited in areas where security clearance is not a problem. The following is a detailed description of the 1998 air monitoring program active sites.

Site No. 4: The sampling trailer is located about 210 feet due north and 70 feet west of Firing Site 16. The equipment in the trailer includes a weather system that measures wind direction, wind speed, temperature, and relative humidity; the VOC sampler; and a data acquisition system. TSP and PM-10 samplers are placed upon raised platforms near the trailer on a reinforced concrete pad.

Site No. 5: The sampling trailer is located just inside the perimeter fence on Farm-to-Market Road 293, approximately 3,400 feet northeast of the burning grounds, at the fourth power pole approximately 600 feet west of a farmhouse. The farmhouse is the closest residence near the Pantex facility. Sampling equipment at this site is an exact replica of that at Site No. 4, except that a colocated PM-10 sampler for quality assurance sampling is located at Site No. 5, rather than a colocated TSP sampler, and two colocated VOC samplers are located in the trailer.

Site No. 7: The sampling trailer is located at the Masterson pump station, northeast of the Burning Grounds. The trailer contains an FTIR detector with dual 100-meter White cells. A meteorological station is integrated with the FTIR to incorporate wind direction, atmospheric pressure, and temperature measurements into the sample data files. Because of the higher detection levels associated with measurements made using this method, these measurements are classified as non-critical, taken for informational purposes only.

Sampling for TSP, PM-10, and VOC's is on an episodic basis. Each sample is collected for a period of 24 hours. The FTIR samples are taken every six minutes, alternating between a scrubbed (of water and carbon dioxide interferents) and unscrubbed cell. Meteorological data are collected continuously with a datalogger recording five-minute converted averages that are eventually converted into hourly averages. All noncontinuous pollutant samples are transported to the TNRCC's Austin laboratory and analyzed.

IV.C.3.e Biota/Foodchain

Surveillance of plant and animal life is important to understanding and assessing the effects of Pantex Plant operations on the environment. Vegetation samples taken for radionuclide and fluoride analyses represent both native species and crops (see IV.C.4.g for discussion of radionuclide sampling results). Native vegetation on the Southern High Plains is primarily range grasses and forbs (weeds). Crops are defined as any agricultural product that can be harvested or gathered for food, forage, or fiber. Forage, often referred to as fodder, is plant matter composed of leaves and stems that is used as food by domestic animals. Because various species and individuals take up contaminants differently under different growing conditions, data interpretation is complex, and results must be considered and evaluated in concert with results of surveillance of other media, such as soils and air.

Systematic faunal sampling for radiological contamination was initiated in July 1995. Animals at Pantex Plant are sampled to determine whether the Plant's activities are having an impact on populations. Prairie dogs are the main species sampled at the Plant because they interact with other environmental media being analyzed; they eat vegetation, dig in the soil, and breathe soil particulates daily. Prairie dogs are numerous enough that sampling will not seriously impact their population, and samples can be collected from the same general location repeatedly. Because no prior systematic data are available for comparison, the 1995 data serve as a baseline for future monitoring.

The only non-radiological parameter currently under surveillance is inorganic fluoride concentrations in vegetation. Pantex Plant's current operations include thermal treatment of explosives and explosive contaminated materials at the Burning Grounds, which may release fluoride particulates to the atmosphere that can be taken up through microscopic openings in the leaves of vegetation. Some animals, particularly cattle, are more sensitive than people to fluoride and may develop fluorosis from grazing on contaminated vegetation. The Texas Air Control Board, a predecessor agency of the Texas Natural Resource Conservation Commission (TNRCC), established limits for inorganic fluoride in air and forage in Texas Administrative Code, Title 30, Chapter 113, "Control of Air Pollution from Toxic Materials." For 1995, this limit varies from 20 to 60 parts per million (ppm) by weight, depending on the number of consecutive monthly samples analyzed from any location.

Winter wheat, grain sorghum, forage sorghum, and native vegetation are grazed by cattle at Pantex Plant; therefore samples of these plants were analyzed for inorganic fluoride. Samples from crops and native plants were collected in accordance with plant procedures for inorganic fluoride in vegetation. The fluoride sampling was begun in March 1995. Samples were collected and analyzed from at least five on-site and two off-site control locations each month except May. One of the five on-site samples was always scheduled to be taken from Playa #3, which was not sampled if vegetation was submerged, as it was in May. All sampled locations with detectable fluoride were below the regulatory limits.

IV.C.4 Media-Specific Contamination (Radiological)

IV.C.4.a Definitions of Radiological Terminology and Introduction

The following terms are used in this section:

AEC (Atomic Energy Commission) The AEC was established after the end of World War II to direct the uses of nuclear energy. The AEC was dissolved in the early 1970s into the Nuclear Regulatory Commission (NRC) and Energy Research and Development Administration (ERDA). The NRC was responsible for regulating nuclear energy, power, and uses of radioactive materials. ERDA was responsible for nuclear weapons production. Later ERDA's name was changed to the Department of Energy.

Alpha Particle A charged particle emitted from the nucleus of an atom having a mass and electrical charge (+2) equal to that of a helium nucleus: i.e., two protons and two neutrons.

Beta Particle A charged particle emitted from the nucleus of an atom. Each beta particle has a -1 electrical charge and a mass equal to an electron. Once emitted, there is no difference between the beta particle and the electron.

DCG (Derived Concentration Guide) The concentration of a radionuclide in air or water that, under conditions of continuous exposure for one year by one exposure mode (either ingestion of water, submersion in air, or inhalation), would result in an effective dose equivalent of 100 mrem. DCGs do not consider decay products when the parent radionuclide is the cause of the exposure.

Dose Equivalent The radiation dose arising from an exposure to the whole body or parts of the body modified using radiation protection factors that take into account the type of radiation (alpha or beta particles; radiation weighting factors) and the dose to tissues (tissue weighting factors). The radiation weighting factors describe the relative damage caused by the types of the radiation. The tissue weighting factors describe the sensitivity of different tissues to radiation.

Gross Alpha and Gross Beta Radiation Measurements Estimations of the total amount of alpha or beta emitting radionuclides in the environmental media. The measurement can only serve as an approximation or starting point for further analyses. Because these types of measurements do not identify the radionuclide(s), it is weighted relative to the radionuclide selected for instrument calibration; no estimate of radiation dose can be made from these values.

Maximum Contaminant Level (MCL) The MCL is the regulatory limit for contaminants in drinking water supplies. For radioactive materials, the MCL, which varies with the radionuclide, is based on the radiation dose delivered to an individual who consumes 2 liters of water per day every day of the year.

NRC (Nuclear Regulatory Commission) An independent agency established by the U.S. Congress under the Energy Reorganization Act of 1974. The NRC mission is to ensure adequate protection of the public health and safety, the common defense and security, and the environment in the use of nuclear materials in the United States. The NRC's scope of responsibility includes regulation of commercial nuclear power reactors; nonpower research, test, and training reactors, fuel cycle facilities; medical, academic, and industrial uses of nuclear materials the transport, storage, and disposal of nuclear materials and waste. NRC regulations do not specifically apply to DOE facilities such as Pantex but comparison of NRC compliance values and DOE values can be used as a reference point.

Radioactivity Units Radioactivity units are used to quantify radioactive material. The units are based on the rate at which radioactive materials decay; that is, the number of transformations (atomic disintegrations) per second (dps). The units are expressed in two systems, a conventional system and the Systeme International (SI). A curie (Ci) of radioactive material is defined as 37 billion dps; one becquerel (Bq) is defined as one dps. One Bq equals 27 x 10^-12 Ci. Because of the small amounts of contaminants found in the environment at Pantex, prefixes are used. Table 1 lists the conventional and SI units.

Radiation Dose The radiation dose is a generic term for any of the following terms. Specifically, the absorbed dose describes the amount of energy deposited by radiation into a system of a given mass. The units are the rad or the gray (Gy). The effective dose adjusts the absorbed dose for the type of radiation (alpha, beta, gamma). The equivalent dose weights the effective dose with a factor for the type of tissue irradiated. The units are the rem or the sievert (Sv). As with the radioactivity units, conventional units and SI units are used. The conventional system uses the rad and rem and the SI system uses the gray and the sievert. The same prefixes discussed in the radioactivity units are used for the radiation dose. There are 100 rads or rem in one Gy or Sv, respectively. The average background radiation exposure in the United States is 3 mSv (300 millirem) which includes radon.

Radioactivity per Mass Unit The radioactivity of a material divided by a mass. This is usually expressed as activity per gram (1/30^th of an ounce) or per kilogram (2.2 pounds per kilogram). An example of such usage is pCi per gram (g).

Radioactivity per Volume Unit Expressed as amounts of radioactivity per cubic meter or milliliter of water or air. A cubic meter contains over 264 gallons; whereas, a milliliter (mL) equals about 1/30^th of one ounce. A liter is 0.946 quarts or 30 ounces.

RCG (Radiation Concentration Guides) Established by the AEC and were used to set limits of exposure to workers or members of the public. The RCG represents the dose limit delivered by a specific amount of radioactivity in a specific environmental media. For the public, the RCG was 170 millirem.

SI (Systeme International or International System) is a relatively new single system designed to provide a unified system of technical terminology within the scientific community. In this report, the SI units will be given with the conventional units in parenthesis.

Thermoluminescent Dosimeters (TLD) TLDs are used to monitor long-term levels of gamma or photon radiation in environmental settings. TLDs are also used for personal dosimeters. The most common types of TLDs are made of lithium fluoride.

Uranium, Depleted A product of the uranium enrichment process, depleted uranium (DU), is composed of mostly U-238 (minimum 99.3% U-238). The amounts of U-234 and U-235 present in natural uranium have been artificially separated. The DU is used in industry and the military for uses such as ballast in aircraft and elevator counterbalances, radiation shielding, armor, and armor piercing shells. The DU can be used in nuclear weapons.

Uranium, Naturally Occurring Natural uranium is composed of 3 chemically identical forms or isotopes that occur at different abundances. They have the same number of protons and electrons but the number of neutrons differ. The atomic mass is the sum of the protons and neutrons. In nature, the isotopes of uranium with their abundances are U-238 (99.275%), U-235 (0.72%) and U-234 (0.0055%). Of the three isotopes, only U-235 can be used as nuclear weapons material and it must be separated through a series of complex technical steps. In Texas, the apparent background concentration of total uranium in Pullman soils which compose much of the Panhandle area ranges from 0.034 to 0.046 Bq/g (0.91 - 1.25 pCi/g). The background levels of uranium collected near the Bushland location were reported at about 0.031 Bq/g (0.84 pCi/g).

The Pantex Plant Annual Environmental Monitoring Reports were reviewed for the years covering 1973 through 1996 as were the quality assurance and quality control program reports for radiological sampling and laboratory analysis. Many of these reports were prepared prior to the remedial investigations at the site. Those reports contain descriptions of the radiological monitoring and sampling programs in support of environmental restoration and protection around the Pantex plant. Pantex monitors surface waters, groundwater, soils, air, and biota in and around the facility. Another data source used was the USGS National Geochemical Data base: National Uranium Resource Evaluation Data for the Conterminous United States, USGS Digital Data Series DDS-18-A (1994). The following is a summary of these environmental reports.

In general, Pantex initiated environmental monitoring in 1972 when they began air, water and vegetation monitoring. At that time, the monitoring reports contained a limited amount of data. Since then, the number of samples has increased on a yearly basis and the range of the sample area has increased. For example, in 1974, Pantex added tritiated water (HTO) monitoring to the air monitoring system and 7 rabbit specimens were collected. Today non-radiologicals and more plant and animal species are included in their reports. In the early reports, the data were expressed in units commonly used at that time. To compare those data with current terminology required conversion to the more commonly used terms of today.

The Los Alamos National Laboratory estimated the annual releases of depleted uranium (DU) from 1963 to 1979. These estimates are U-238 (595 million Bq; 16.1 mCi), U-234 (222 million Bq; 6 mCi), and U-235 (11.1 million BQ; 0.3 mCi). Aerial surveys located the majority of the DU (96%) around Firing Site 5 with the remainder around Firing Site 4.(8)

IV.C.4.b Groundwater

The waters sampled in the 1990s included groundwater from the Ogallala aquifer, production wells, and monitoring wells. Radioactivity in groundwater samples or wells are below the current or proposed Maximum Contaminant Levels (MCL); that is, the radiological dose is expected to be less than 0.04 mSv (4 mrem) per year. Appendix A contains the list of the radionuclides monitored or analyzed for in the groundwater.

IV.C.4.c Surface Waters

Waters contained on-site in some playas were higher in radioactivity than off-site background levels; however, these were still below MCL values. Few surface water samples were collected off-site in the early years of environmental sampling. In 1975 and 1976, elevated levels of gross alpha radioactivity (concentration) were found at stations WS 1 (near Pantex Lake), 2, and 6. However, no conclusions can be drawn from these samples. In the 1980s, water samples collected were all below the levels mandated in the Safe Drinking Water Act.

IV.C.4.d Soils

The radioactivity concentrations in soil samples during 1973 were less than the established RCG for members of the public. Based on a percent of the RCG, the highest value was less than 10% of the RCG and that was plutonium (Pu), presumably from world-wide atmospheric testing. The AEC (now DOE) did share the environmental samples with the EPA but at that time, the comparisons of the AEC results with EPA results were not very good. The average uranium in soils was about 0.0056 Bq/g (0.15 pCi/g) and plutonium was about 0.015 Bq/g (0.4 pCi/g). The EPA values were higher for plutonium but lower for uranium - but still below levels that would cause adverse health effects.

In 1975, elevated levels of uranium were found in the soils north of Zone 4; but, the values were still less than 0.11 Bq/g (3 pCi/g), about 3 times above background levels. Radioactivity concentrations in soils of the 1990s show trends similar to that in waters. The radioactivity levels in off-site soils are similar or equal to background levels both in the vicinity of Pantex and at other locations in the United States. However, uranium levels in soils of several Pantex firing sites (Sites 4 and 10) strongly suggest the presence of depleted uranium. Elevated levels of uranium in soils were reported at several firing site sampling locations, FA-SS-01, FB-SS-05 through 08, FC-SS-04 and 05. The levels, however, are within acceptable health and exposure levels of both the DOE and the Nuclear Regulatory Commission for areas of controlled access.

In 1985, no plutonium was detected in either on-site or off-site soils. Uranium in soils near the site (off-site) were comparable to background locations. High levels of uranium (several times above background) were still detected in soils at several firing site sampling locations.

In the 1986 environmental report, off-site sample location OS-SS-06 showed elevated uranium in soil 0.48 Bq/g (13 pCi/g). This sample was collected north of Zone 4 and appears to have been collected at the fence line (property boundary). Soil samples around Firing Sites 4 and 5 continued to show high uranium in soils.

Pantex did not report releases of uranium in 1987; however, samples OS-SS-04 and samples OS-SS-01 (NW), 02 (NWN), 10 (N Zone 4), 18 (NE Pantex Lake), 30 (NW of lake), and 31 (SW of lake) showed high uranium in soil. Firing site samples FA-SS-01 and FB-SS-01 continued to show very high uranium in soils.

More recent reports cover the Pantex area in more detail with respect to uranium and plutonium concentrations in soils both on-site and off-DOE property. In all cases of off-site samples, there were no levels of uranium or plutonium found that would be considered a health hazard.

Soil samples from Landfill 6 located within Solid Waste Management Unit 57 show elevated levels of uranium in soil averaging 0.38 Bq/g (10.25 pCi/g) in both landfill debris and in subsurface soils outside the landfill proper. The method used for this evaluation was a gamma scan (nonspecific) and when samples were re-analyzed using alpha spectroscopy (specific analysis), U-238 was found at background levels.(9) In general most samples for this particular site were qualified estimates as contamination was also present in the blanks prepared for quality control. The actual concentrations could be higher or lower than reported because uranium was present in the quality control samples.

Soil concentrations of uranium in Firing Sites 5, 6, and 15 (ADS 1205) show that thorium-232 (Th-232) and one of its decay products, Th-228, appear to be in equilibrium from 0 to 2 feet in depth. However, the relative concentrations of two uranium isotopes, U-238/234, indicate the presence of depleted uranium as expected from firing site activities. Depleted uranium was found in excess of 1.1 Bq/g (30 pCi/g) at 6 locations from the surface to 2 feet in depth. The range found was 1.3 - 19.4 Bq/g (35.8 - 523.5 pCi/g). Very little U-235 was found in these samples, 0.033 Bq/g (0.89 pCi/g) and 0.28 Bq/g (7.5 pCi/g). The highest reported value was 1-2 feet down and located near the foot of a berm. At the same location, DU was found at concentrations ranging from 1.3 - 5.8 Bq/g (34 - 156 pCi/g) at 3, 4 and 5 feet (9).

Within the area of the Nuclear Weapons Accident Residue site, some indication of DU existed. The Nuclear Weapons Accident Residue site contained debris from weapons accidents. All debris had been removed by the late 1980's to other DOE complexes. The expected ratio of U-238/U-234 should have been 1.0 but at this site the ratio was 6.5. However, the total activities still were less than 1.1 Bq/g (30 pCi/g) (10) which is not considered a health hazard.

IV.C.4.e Ambient Radiation

Ambient radiation is photon radiation present in the environment. This type of radiation, measured by thermoluminescent dosimeters (TLD), is statistically indistinguishable from on-site background levels. Ambient levels of radiation as detected by TLD readings beginning in 1985 were similar both on and off-site, that is they were comparable to background levels.

The current TLD readings show an average annual exposure of approximately 1 mGy (100 millirad) per year both on-site and off-site. This value is also consistent with ambient exposures across the country at similar altitudes.

IV.C.4.f Air

Pantex established a system of air monitoring stations consisting of 9 continuous air monitoring stations at a 5 mile radius in November 1973. Early Pantex reports estimated atmospheric releases of depleted uranium (DU) and tritium (H-3) but stated that there had been no releases of plutonium (Pu). In 1973, the amount of DU released was 7.4 billion Bq (0.2 Ci) and the amount of tritium released was 3.7 billion Bq (0.1 Ci). Much of the data were reported as an annual summary; however, the DU information can be misleading if all releases occurred as a result of brief periods of testing at the firing sites.

Pantex air releases in 1974 were similar to the previous year. The estimated doses at the fence line were not at levels known to cause adverse health effects.

In 1982, more tritium was released to the atmosphere than in previous years. In 1982, the maximum levels of tritium in air ranged from 0.059 - 0.24 Bq/m³ (1.6 to 6.4 pCi/m³). Although more tritium than usual was released, the estimated dose to the public was less than 0.01 mSv (1 mrem); much less than the 3 mSv (300 mrem) radiation exposure received from normal background (natural radiation sources).

In 1985, elevated levels of tritium were detected off-site; however, the levels were still below levels known to cause adverse health effects and, for comparison, below levels of NRC compliance.

In 1989, a major release of tritium occurred from within the assembly area. Dose assessments developed by DOE and re-evaluated by Battelle in 1994 suggested that tritium doses were well below levels known to cause adverse health effects. As late as 1994, tritium continues to be released from the gravel dome covering the release area and soil resulting in continuously decreasing doses (Table 2). The Battelle reassessment of the tritium release model used indicates that doses from 1990 to 1992 ranged from 4 to 70 x 10^-6 mrem.

In these dose assessments, the greatest contributor to dose was fruits and nonleafy vegetables - future public health actions should (if release occurs again) be taken during the growing season.

Air monitoring for radionuclides (U, Pu, and tritium) in the 1990s shows the atmospheric concentrations of the U and Pu radionuclides to be at levels similar or equal to background levels and would not cause adverse health effects. In comparing the air concentrations around Pantex to other cities in Texas (Austin and El Paso), the air actually contains less radiological contaminants than those two cities. Tritium (H-3) concentrations have been elevated around Pantex in the past as a result of a major release occurring in May 1989. The elevated levels of tritium, however, have not resulted in radiation exposures or doses above levels that would cause adverse health effects.

Table 2 Tritium Releases and Associated Doses
Year	Tritium Released	Estimated Dose to the Maximally Exposed Individual
1989	1.48 x 1015 Bq (40,000 Ci; accidental release)	14.3 µSv (1.43 mrem) (assumes uniform release during year)
1990	9.44 x 1013 Bq (2550 Ci)	1.6 µSv (0.16 mrem) (assumes uniform over year - trapped tritium)
1991	6.29 x 109 Bq (0.17 Ci)	1 x 10-10 Sv (1 x 10 -5 mrem) (no major releases)
1992	4.66 x 109 Bq (0.126 Ci)	0.9 x 10-10 Sv (9 x 10 -6 mrem) (no major releases)

IV.C.4.g Biota/Foodchain

Vegetation samples collected and analyzed for tritium suggest that off-site tritium in vegetation is higher than samples collected on-site. The highest values were from samples collected south-southeast to west of the plant. The predominant wind direction is from the south. However, during the 1989 tritium release, the wind was from the opposite direction and was followed by a thunderstorm. Pantex reported the maximum dose as a result of the release was toward the east-southeast direction.

Although the levels of tritium in vegetation may be elevated above the on-site values, the levels are well below levels known to cause adverse health effects. For example, an individual would have to consume over 5,000 tons of affected vegetation per year to approach a dose knowned to cause adverse health effects.

In the early years of biota sampling at Pantex, similar results for uranium were seen. Very few samples showed elevated levels of uranium in foodstuffs and these levels were below levels that adversely affect health.

IV.D Quality Assurance and Quality Control

ATSDR obtained Quality Assurance/Quality Control (QA/QC) summaries from EPA and DOE with respect to environmental monitoring data. No analytical problems were noted in the QA/QC summaries obtained.

IV.E Physical and Other Hazards

No accessible physical hazards were identified at the site. Protective measures for site workers are included in specific workplans.

V. PATHWAYS ANALYSIS

To determine whether people are exposed to contaminants migrating from a site, Agency for Toxic Substances and Disease Registry (ATSDR) representatives evaluate the environmental and human components leading to human exposure. An exposure pathway consists of five components: (1) a source of contamination, such as drums or waste pits; (2) an environmental medium in which the contaminants might be present or from which they might migrate, such as groundwater or soil; (3) points of human exposure, such as drinking water wells or work areas; (4) routes of exposure, such as inhalation, ingestion, or dermal absorption; and (5) a potentially exposed population.

Although historical Pantex Plant operations have resulted in contamination of on-site perched groundwater, surface water, soil, sediment, and air, no on-site potential or completed exposure pathways are evident from Pantex Plant activities. Because of the contaminants' isolated locations and security measures at the storage and disposal areas, humans have little opportunity for contact with contamination on site.

The contamination of a downgradient Ogallala drinking water well of a private residence was evaluated as a potential off-site exposure pathway. The potential pathway was eliminated because the old well was closed and sealed, and the new well was constructed in such a manner as to prevent intrusion of the contaminated perched aquifer. In addition, a particulate and carbon filtration system sampled and maintained by Pantex Plant staff was installed on the new drinking water well.

Tritium concentrations in air have been elevated in the past as a result of a major release occurring in May 1989. The elevated air levels of tritium, however, have not resulted in radiation exposures or doses above levels that would cause adverse health effects. Although the levels of tritium in off-site vegetation are elevated above the on-site values, the levels are well below levels that would pose a health threat. For example, an individual would have to consume more than 5,000 tons of affected vegetation per year to approach a dose that would cause adverse health effects.

Radioactivity in the Ogallala aquifer, which is the source of the water supply for the plant and for the City of Amarillo, is well below the current or proposed Maximum Contaminant Levels and, therefore, is not a health hazard.

Chemical contamination above health screening values has been detected in the perched aquifer at the southeastern plant boundary; however, that particular perched aquifer is not used as a water source downgradient from the contaminant plume, so no one is exposed to the contaminants. If the aquifer were used in the future as a water source, water from that perched aquifer would require treatment to remove contamination before it would be safe for consumption.

VI. PUBLIC HEALTH IMPLICATIONS

VI.A Toxicological Evaluation

A release of a hazardous waste does not always result in exposure. People are exposed to a contaminant such as those identified at the Pantex Plant Site only if they come in contact with it; they might be exposed by breathing, eating, or drinking a substance containing the contaminant or by skin contact with a substance containing the contaminant. Several factors determine the type and severity of health effects associated with exposure to a contaminant. Such factors include the exposure concentration (how much); the frequency and/or duration of exposure (how long); the route of exposure (breathing, eating, drinking, or skin contact); and the multiplicity of exposure (combination of contaminants). Moreover, people can be exposed to an environmental contaminant by more than one route of exposure. Once exposure takes place, characteristics such as age, sex, nutritional status, genetics, lifestyle, and health status of the exposed individual influence how the individual absorbs, distributes, metabolizes, and excretes the contaminant. Together, those factors and characteristics determine the health effects that might result from exposure to a contaminant.

The Agency for Toxic Substances and Disease Registry (ATSDR) considers the previously described physical and biologic characteristics when developing health guidelines. Toxicological profiles prepared by ATSDR representatives summarize chemical-specific toxicologic and adverse health effects information. Health guidelines, such as ATSDR's minimal risk level (MRL) and the Environmental Protection Agency's (EPA) reference dose (RfD) and cancer slope factor (CSF) are included in the toxicological profiles. Those guidelines are used by ATSDR public health professionals to determine an individual's potential for developing adverse noncancer health effects and/or cancer from exposure to a hazardous substance.

Health guidelines provide a basis for comparing estimated exposures with concentrations of contaminants in different environmental media (soil, air, water, and food), depending on the characteristics of the people who might be exposed and the length of the exposure. An MRL is defined as an estimate of the daily human exposure to a contaminant that is likely to be without an appreciable risk of adverse noncancer health effects over a specified duration of exposure (acute, less than 15 days; intermediate, 15-365 days; chronic, greater than 365 days). Oral MRLs are expressed in units of milligrams per kilogram per day (mg/kg/day). MRLs are not derived for dermal exposure. The method for deriving MRLs does not include information about cancer; therefore, an MRL does not imply anything about the presence, absence, or level of cancer risk. An EPA RfD is an estimate of the daily exposure of the human population, including sensitive subpopulations, that is likely to be without appreciable risk of adverse noncancer health effects during a lifetime (70 years). Noncancer health guidelines are adjusted downward using uncertainty factors to make the guidelines adequately protective of the public health. Therefore, the health guidelines should not be viewed as strict scientific boundaries between what level is toxic and what level is nontoxic. For cancer-causing substances, EPA has established the CSF as a health guideline. The CSF is used to estimate the number of excess cancers maximally expected from exposure to a contaminant.

To link a site's human exposure potential with health effects that might occur under site-specific conditions, ATSDR representatives estimate human exposure to site contaminants from ingestion and/or inhalation of different environmental media. The following relationship is used to determine the estimated exposure to the site contaminant:

ATSDR uses standard intake rates for ingestion of water and soil. The intake rate for drinking water is 2 liters per day (L/day) for adults and 1 L/day for children. For incidental ingestion of soil, the intake rate is 100 mg/day for adults, 200 mg/day for children, and 5,000 mg/day for children with pica behavior (repeated ingestion of nonnutritive substances). Standard body weights for adults and children are 70 kg and 10 kg, respectively. The maximum contaminant concentration detected in a specific medium at a site is used to determine the estimated exposure; use of the maximum concentration results in an evaluation that is most protective of human health. When unknown, the biological absorption from environmental media (soil, water, etc.) is assumed to be 100%.

For radiological contaminants, ATSDR uses information on radiation exposure and its effects related to environmental levels prepared by federal agencies, including EPA, Department of Energy (DOE), and the Nuclear Regulatory Commission (NRC). The agency also uses other publicly available data sources and recommendations on radiation dose limits. The National Council on Radiation Protection and Measurements (NCRP), the International Commission on Radiological Protection (ICRP), and the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) develop these sources.

Pantex Plant on-site contamination is not considered to pose a health hazard. It is highly unlikely that the general public contacted (past scenario), is contacting (current scenario), or will contact (future scenario) on-site contaminated environmental media. Off-site migration of contamination has been identified in the perched aquifer at the southeastern plant boundary; however, that particular perched aquifer is not currently used as a water source in the direction of contaminant migration, and the sole downgradient private drinking water well has been replaced with an upgraded new well with a particulate and carbon filtration system. Contamination above health screening levels has not been detected in the regional source for drinking water, the Ogallala Aquifer.

Because the individual contaminants detected at this site are present at levels not expected to result in adverse health effects, ATSDR considers the combined effect of these contaminants as not likely to be of public health concern either. Research has shown that a mixture produces no adverse health effects in dosed animals when the components of that mixture are present at levels below their respective No Observed Adverse Effect Levels (NOAEL); for example, at concentrations that would have caused no adverse effects in animals treated separately with those component chemicals. Considering that ATSDR comparison values are typically 100 to 1000 times lower than the corresponding NOAELS, it is reasonable to expect that environmental contaminants will not produce any combined effects, even if their individual concentrations exceed their respective ATSDR comparison values. Therefore, there are no adverse toxicological implications to Pantex Plant off-site residents.

VI.B Health Outcome Data Sources

To address community health concerns regarding the occurrence of certain health conditions in this area, the Texas Department of Health (TDH) evaluated data from several existing State of Texas vital records and disease surveillance databases. The databases used are described in this section. Depending upon availability, incidence data (new cases) and/or mortality data (deaths) from the four counties closest to the Pantex Plant (Armstrong, Carson, Potter, and Randall) were evaluated; however, there are problems associated with the interpretation of vital record and disease surveillance data. A discussion of these problems is presented in the appropriate sections of this report. An evaluation of the data and answers to questions posed by the community are also provided in subsequent sections.

VI.B.1 Birth Defects

All data relating to birth defects were provided by the Texas Department of Health Birth Defects Monitoring Division (TBDMD). The TBDMD identified cases of birth defects by examining three types of vital record certificates:

1) Live birth certificates
2) Fetal death certificates
3) Infant death certificates

These certificates, which are available from the Texas Department of Health Bureau of Vital Statistics (BVS), contain specific information on birth defects and cause of death. When a child is born with a birth defect, attending health care personnel may indicate the presence of a birth defect by checking the appropriate category of birth defect on the birth certificate (Appendix E; Table 1). Fetal death certificates also have these check boxes. Additionally, they list the International Classification of Disease 9th Revision (ICD-9) code for the underlying cause of death. Infant death certificates list the ICD-9 codes for all causes of death (underlying cause and others). For birth certificates, birth defects listed in the check boxes were counted. For fetal death certificates, the birth defects listed in the check boxes and the birth defects listed in the underlying cause of death section of the certificate were counted. For infant death certificates, all birth defects in both the underlying and all causes of death sections of the certificate were counted.

Case Definitions

For this health assessment, the TBDMD defined a case as an infant or fetus who meets all of the following criteria:

1) Was delivered between January 1, 1990 and December 31, 1994;
2) Had a mother residing either in Armstrong, Carson, Potter or Randall counties at the time of delivery; and
3) Had a birth defect indicated on the vital record.

Because of a change in the type of information required to be reported on birth certificates in 1989, the TBDMD considers 1990 as the first year for which reliable data on specific birth defects are available. The last year for which complete data are available is 1994. In addition to analyzing the data for each of the individual four counties, the TBDMD also analyzed the data for Potter and Randall counties combined, and for all four counties combined.

Limitations With These Types of Data

A brief discussion of the limitations associated with these types of data is warranted. In general, analyzing vital record data for many different birth defects may result in the detection of a number of apparent random clusters of birth defects in time. Often it is difficult, if not impossible, to distinguish these random clusters from clusters of cases which may be caused by common environmental factors. In addition to this random statistical variation there is a potential problem with variations in case ascertainment and case definition. Differences in medical practice, recognition of syndromes, and changes in autopsy services alone often can explain many variations over time in the frequency of specific defects.

Birth Certificates

Data obtained from live birth certificates are limited by the nature of the document. In Texas, birth certificates are required by law to be filed with the State within five days of the birth. The expeditious nature of the filing system is a potential source for reporting bias. For instance, if a birth defect is not readily apparent or additional medical tests are required to confirm a specific diagnosis, the birth defect may not be detected prior to the filing of the birth certificate. Reporting of birth defects on birth certificates also may not be uniform throughout the State. Hospitals or birthing centers that are more fastidious about reporting birth defects on birth certificates would have better reporting than the State as a whole.

Fetal Death Certificates

A fetal death certificate is required for any stillborn fetus delivered after 19 weeks gestation. Similar to the recording of birth defects on birth certificates, fetal death certificates also are prone to reporting bias. Depending on the gestational age of the fetus, often it is difficult to determine whether the fetus died from a birth defect. This issue is compounded by the fact that autopsies are less likely to be performed on fetuses. The diagnosis of cause of death may be more vague because the incomplete formation of organs may preclude the detection of the birth defect. As with birth certificates, reporting of birth defects on fetal death certificates may not be uniform throughout the State. Hospitals that are more fastidious about reporting defects on the fetal death certificates would have better reporting than the State as a whole.

VI.B.2 Low Birthweight

The Texas Department of Health obtained data relating to low birthweight infants from the information recorded on birth certificates. A low birthweight infant is defined as an infant who is born weighing less than 2,500 grams (5.5 pounds). The incidence of low birthweight for the period 1990-1994 was determined. A case was defined as an infant weighing less than 2,500 grams at birth who was born to a mother residing in one of the four counties of interest. There are many factors that can affect the birthweight of an infant. These factors are discussed in section VI.C.2 (Low Birthweight) of this report.

VI.B.3 Cancer

All cancer data were provided by the Texas Department of Health Texas Cancer Registry (TCR). The TCR maintains cancer incidence and mortality data for Texas. The TCR obtains mortality data from death certificate information maintained by the Texas Department of Health Bureau of Vital Statistics (BVS). Cancer incidence data are acquired under the Texas Cancer Incidence Reporting Act (Chapter 82, Health and Safety Code) which requires every general and special hospital, clinical laboratory, and cancer treatment center to report all cases of cancer to the TCR. Every inpatient or outpatient case diagnosed with or treated for cancer must be reported to the TCR. Although the TCR is a passive registry that relies on facilities to supply the information, it monitors the number of expected reports from each institution and contacts those facilities that fail to report. To ensure that reported data are complete and accurate, TCR staff perform case-finding and other quality control checks at these institutions. The TCR has determined that for Public Health Region 1, which includes the counties of concern, cancer incidence reporting is complete for the years 1985-1993.

Incidence Data

The TCR evaluated cancer incidence data for Armstrong, Carson, Potter, and Randall counties for the period 1985-1993. Specifically, the TCR evaluated incidence data for the following cancer sites (types of cancer): leukemia, cancer of the lung, bone, prostate, breast, brain, thyroid, and all cancer sites combined. Cancer incidence data were analyzed for Armstrong and Carson counties separately and for Potter and Randall counties combined. The cancer incidence data for Potter and Randall counties were combined because Amarillo is located in more than one county and zip codes cross county lines; therefore, a cancer diagnosis may not always be coded to the correct county.

To maintain consistency with a previous investigation (11) of cancer incidence data for this area, TCR also analyzed data for the four major cell-specific types of leukemia; acute lymphocytic leukemia (ALL), chronic lymphocytic leukemia (CLL), acute myelocytic leukemia (AML), and chronic myelocytic leukemia (CML) for the period 1985-1993.

Mortality Data

Using death certificate information, the TCR also evaluated cancer mortality for Armstrong, Carson, Potter, and Randall counties for the period 1985-1994. Specifically, the TCR evaluated mortality data for leukemia, cancer of the lung, bone, prostate, breast, brain, thyroid, and all cancer sites combined. Cancer mortality data were evaluated for each of the four counties separately because the procedural anomaly only occurs with the coding of cancer incidence cases and does not affect cancer mortality data.

One of the limitations with using computerized death certificate cancer mortality data is that only the underlying or primary cause of death is available. Because people may live for many years with cancer and die from some cause not related to the cancer, there is a potential for an under reporting of these specific cancers. Any mortality data analysis suffers from the loss of people contracting cancer but who moved away from the region before they died, as well as people who moved into the region with pre-existing cancer.

VI.B.4 Muscular Dystrophy, Multiple Sclerosis, Amyotrophic Lateral Sclerosis, and Lupus Erythematosus

In Texas, muscular dystrophy, multiple sclerosis, amyotrophic lateral sclerosis, and lupus erythematosus are not reportable diseases; therefore, incidence data are not available. However, mortality data for these diseases in each of the four counties and in Potter and Randall counties combined are available for the period 1980-1994 from death certificate information provided by the Texas Department of Health Bureau of Vital Statistics.

As with cancer mortality data, there also are limitations with computerized death certificate mortality data for other chronic diseases which may result in differences in the reporting of these conditions. Diagnosis of these conditions may be difficult and may depend on the diagnosing physician. This could effectively reduce our confidence in the results when comparing rates from one area to another. This issue is further addressed in Appendix F.

VI.C Health Outcome Data Evaluation

For all health conditions examined below, the observed occurrence was compared to the "expected" occurrence using ratios. The expected occurrence was based on the occurrence observed in a reference population, usually the State of Texas as a whole. In some instances State of Texas data were not available; thus, other reference populations were used. For birth defects and low birthweight babies, the number of cases observed was divided by the number expected, producing an observed:expected ratio or OER. For cancer and chronic diseases, the ratio of observed to expected number of cases (incidence) or deaths (mortality) was examined and the information was further standardized to eliminate possible effects due to race, sex, and age. These ratios are referred to as the standardized incidence ratio (SIR) or standardized mortality ratio (SMR).

When any of the above ratios is equal to one (1.00), then the number of observed cases/deaths is equal to the number of cases/deaths that would have been expected based on the incidence/mortality experience of the reference population. When the ratio is less than 1.00, fewer people have the disease or have died from the disease than would have been expected. Conversely, a ratio greater than 1.00 may indicate that more people have the disease or have died from the disease than we would have expected.

Because rates and ratios fluctuate from year to year, one can only estimate the underlying ratio. To determine the precision of our estimate, statistics are used to calculate a 95% confidence interval (95% CI). This is an estimate of the range in which one would expect the underlying ratio to fall 95% of the time. If the 95% CI includes 1.00, then we are reasonably certain that the ratio is within the range that would be expected based on the reference population. When the 95% CI excludes 1.00, the ratio is not within the range that would be expected based on the reference population, and is "statistically significant". As defined here, "statistically significant" means that there is less than a five percent chance that the observed difference is merely the result of random fluctuation in the number of observed cases. The precision of our estimate of the underlying ratio varies inversely with the width of the confidence interval. A narrow confidence interval indicates that the estimate of the underlying ratio is precise. Conversely, a wide confidence interval indicates that the estimate of the underlying ratio is not very precise. Confidence intervals were not calculated where there were no observed cases.

Care should be taken when interpreting the meaning of statistical significance. To report that a result is statistically significant merely means that the result was unlikely to have occurred by chance. Specifically, statistical significance means that the likelihood of the result having occurred by chance is small. A statistically significant result does not necessarily mean that the result is significant in the everyday meaning of the word.

Interpreting statistically significant results often is difficult and requires the consideration of several factors. In addition to the magnitude of the observed to expected ratio; the number of observed cases, the number of expected cases, the width of the confidence interval, and the source of the data all must be considered. For example, in the birth defects tables, one observed case of "other central nervous system anomalies" was 55.73 times what we would have expected. However, based on the reference population, the expected number of cases was 0.0179. Because it is not possible to observe a fraction of a case, observing as little as one case with an expected value less than one results in a relatively large ratio (55.73). In this example the confidence interval tells us with 95% confidence that the underlying ratio could be anywhere from 1.41 to 311.26; however, the width of the confidence interval tells us that in this instance our estimate is not very precise. Although the observed to expected ratio is statistically significant, the small number of observed cases and the lack of precision would make it difficult to provide, with any degree of certainty, a meaningful interpretation.

Care also should be taken when interpreting a statistically significant result based on adequate numbers and precision. Other factors besides biology could be responsible. For instance, categories that include the term "other" in their description are not very specific. Many types of defects are included in these categories. Thus, it would be very difficult to tell if a "statistically significant" result was due to slight (non-significant) elevations in many defects or if it is due to larger increases in one or two specific defects. Differences in reporting practices also could account for apparent statistical differences between populations.

VI.C.1 Birth Defects

For each type of vital record, the TBDMD compared the observed number of cases for each category of birth defect to the number that would have been expected. The number of expected cases was calculated using prevalence rates for the entire State from the same vital record source (e.g. county birth certificates compared to birth certificates for all Texas).(12) The TBDMD did not adjust the ratios for race and maternal age; they determined that it would probably have a low impact for the majority of birth defects.

Birth Certificates

The results of analyses based on birth certificate data are presented in Appendix E; Table 2. For several types of birth defects, the observed number of cases exceeded the expected number. In Armstrong County, six of the birth defect categories had OERs greater than one; however, only one, "other central nervous system anomalies", was statistically significant. In Carson County, five of the birth defect categories had OERs greater than one; only one category, "other gastrointestinal anomalies", was statistically significant. In Potter County, 20 of the birth defect categories had OERs greater than one and 11 were statistically significant. Five of the categories that were statistically significant were nonspecific "other" categories. Two of the statistically significant categories had expected values less than one. In Randall County, 20 of the birth defect categories had OERs greater than one and eight were statistically significant. Five of the categories that were statistically significant were nonspecific "other" categories. One of the statistically significant categories was based on an expected value less than one.

When all four counties were combined, OERs were statistically significant in 12 of the 23 birth defect categories. Five of these 12 categories were the nonspecific "other" categories. One category, "other congenital anomalies", exhibited significantly fewer cases than expected. Overall, for the four counties combined, the number of infants who had birth defects noted on the birth certificate was 2.16 times more than the number expected in the four counties, and was statistically significant. For reasons described previously, it is difficult to interpret statistical results with respect to the "other" categories because many different defects could be included in each of these categories. However, for individual counties, Potter and Randall counties combined, and all four counties combined there were several clinically clear (do not include "other" in their description) categories that were statistically significant. In some cases, these results are based on five or more observed cases, more than one expected case, and relatively narrow confidence intervals. In these instances, random chance alone may not be responsible for the results.

One possible explanation for these results is that these excesses are a consequence of fastidious hospital reporting practices. There are several reasons which support the role of fastidious reporting in explaining the results for this area. First, approximately three-fourths of all births to residents of Potter and Randall counties take place at Northwest Texas Hospital (NTH). Thus, the reporting practices of that hospital would be expected to strongly influence the overall statistics for birth defects reported for these counties. In the past, concern about high rates of birth defects from the Amarillo area was raised by area residents.(13) Generally, birth defects are under reported on vital records. If previous concern about birth defects resulted in more fastidious reporting, comparing birth defects from this area (obtained from birth defects noted on birth certificates) to an expected number of birth defects based on the State experience may not be appropriate and could result in artificially produced excesses.

In an attempt to rule out more fastidious reporting as a possible explanation for the statistically significant excesses, we compared the data from the four counties to two active surveillance systems (the California Birth Defects Monitoring Program, CBDMP, and the Metropolitan Atlanta Congenital Defects Program, MACDP)[Appendix E; Table 3]. Rates obtained from active surveillance systems allow the most accurate estimation of the expected number of cases for a given area. If birth defects observed in the four counties were higher than the expected number based on rates reported by the active surveillance systems, then reporting bias would not be sufficient to explain the excess. When this comparison was made, no significant excesses were observed; therefore, more complete reporting in this area cannot be ruled out as an explanation for the statistically significant excesses obtained using birth certificate information. Some of the specific birth defect categories did not have equivalent diagnoses among the California or Atlanta registries and, therefore, could not be compared.

The TBDMD was able to examine the birth certificate data with respect to parental occupation and place of employment.(14) Additionally, they were able to look at the distribution of the birth defects by zip code. They did not find anything unusual with respect to parental occupation or workplace. Very few parents of children reported to have been born with birth defects worked at the Pantex Plant. Some of the fathers of affected children worked in meat processing, but this is a common occupation in Amarillo. One zip code, 79107 which extends from the western edge of the Pantex Plant along the Potter County/Carson County border towards Amarillo near the plant, appeared to have significant elevations for several of the birth defect categories (Appendix E; Table 4). Many of these excess cases either were based on observed values less than five, expected values less than one, or nonspecific birth defect categories that include "other" in their description. Reporting practices may explain the statistically significant results for three of the birth defect categories; heart malformations, malformed genitalia, and club foot. The 79107 zip code is in Potter County and had the highest number of reported live births (3,529) for any of the zip code areas studied in the four county area. Because the majority of births to residents of Potter County take place at NTH, fastidious reporting practices cannot be ruled out as a possible explanation. Statistically significant excess for these birth defect categories also was observed in other zip code areas. Microcephalus was elevated in 79121; heart malformations was elevated in 79124, 79108, 79106, 79101, 79102, 79104, 79109, 79103, and 79015; malformed genitalia was elevated in 79110; and club foot was elevated in 79015 and 79068. No clear pattern was seen with respect to increasing ratios and proximity to Pantex.

Fetal Death Certificates

From 1990-1994, in the four counties combined, there were 20 fetal deaths where birth defects were indicated as a cause of death (Appendix E; Table 5). Although "hydrocephalus" shows a statistically significant excess of 40.22 times the expected number in Carson County, this excess is based on one observed case, an expected number of less than one case (0.0249), and an extremely wide 95% confidence interval which barely excludes 1.00 (1.02-223.76). For reasons outlined previously in this report we do not consider this statistically significant result meaningful. In Potter County, based on nine cases, the category "other congenital anomalies" was 2.64 times expected and was statistically significant; however, this is a broad based category with nonspecific diagnoses. As mentioned previously, it is difficult to interpret a nonspecific category that includes "other" in its description.

The portion of the fetal death certificate which lists the "underlying cause of death" also yielded sporadic statistically significant results (Appendix E; Table 6). In Carson County, "other congenital anomalies of the circulatory system" (ICD-9 code 747) was 597.49 times what we would have expected; however, in addition to being a nonspecific category, this result was based on one case, an expected number of cases less than 1.00 (0.0017), and a very wide 95% confidence interval (14.89 to 3,277). Randall County had 2 cases of "other congenital anomalies of the heart" (ICD-9 code 746) which was 9.74 times the expected number of less than one (0.205) with a 95% confidence limit that ranged from 1.18 to 35.17. Although these results are statistically significant, the small number of observed cases, the lack of precision in the estimates, and the nonspecific nature of the categories would make it difficult to provide a meaningful interpretation of these results.

Death Certificates for Infants

The TBDMD examined death certificates for infants (children less than one year old). In Carson County, two cases of major heart defects 'bulbus cordis anomalies' and 'anomalies of cardiac septal closure' (ICD-9 code 745.0-745.9) were found to be statistically significant at 21.84 times the expected number (Appendix E; Table 7). This was the only significant excess or deficit observed among the four counties. Because this finding was based on two cases, an expected number of cases less than one (0.0916), and a relatively wide confidence interval (2.64 - 78.87), we would not interpret these results as indicating an increased risk for these defects in this area.

Discussion

Using birth certificates information, the TBDMD found that several categories of birth defects were significantly elevated in specific counties and in the four counties grouped together. In some cases, the results for individual counties were based on few observed cases, less than one expected case, and relatively wide confidence intervals. In these cases, it would be difficult to attach much importance to findings of statistical significance because our confidence in the precision of the data would be limited. In other cases, the category of defect on the birth certificate was such that many nonspecific diagnoses could be included within the category. It is difficult to interpret the importance of statistically significant findings for categories that can potentially include many different types of birth defects. When such increases are observed it is impossible to determine whether the increase is the result of slight elevations among many types of birth defects or the result of a large increase in one or two birth defects which results in a significant overall increase.

For Potter County, Randall County, Potter and Randall counties combined, and the combined four county area, the TBDMD found a statistically significant excess in the reporting of several specific categories of birth defects. Many of these elevated ratios may be due to factors other than random chance because they are based on a substantial number of observed cases, populations large enough to result in reasonable expected values for comparison, and relatively tight confidence intervals.

Generally, birth defects are under reported on vital records. Thus, using the entire State as the comparison population could result in depressed expected rates. In this area, the majority of all live births take place at Northwest Texas Hospital (NTH). Therefore, the reporting practices of this hospital could greatly influence the observed rates for this area. In an attempt to rule out more complete reporting as the cause, the TBDMD compared the rates observed for this area to two active surveillance systems. Because active surveillance systems provide the most accurate results in collecting birth defects, exceeding the rates reported by these systems would eliminate reporting bias as a possible explanation for these results. When this comparison was made, they did not find any significant differences between the rates for this area and those of the active surveillance systems. Thus, more complete reporting in this area cannot be ruled out as an explanation for the statistically significant excesses obtained using birth certificate information.

Among the 390 comparisons made using fetal and infant death certificate information (Appendix E; Table 5-7), we would have expected 5% or 19 of them to be "statistically significant" by chance alone. Only five comparisons in these analyses showed a significant excess of a specific birth defect, each of those was based on very few cases with expected values less than one. Additionally, there was no consistency regarding the specific type of defect across counties. Consequently, these vital records do not offer compelling evidence of an excess occurrence of birth defects.

Currently, the causes of two-thirds of birth defects remain unknown. Many birth defects have a multifactoral etiology, meaning that a variety of both genetic and environmental (diet, lifestyle, and occupational) factors play a role. For example, microcephaly may be caused by one of several inherited syndromes, by degenerative brain disorder, birth trauma, intrauterine infection, or exposure to radiation in utero. The causes of tracheo-esophageal fistula, cleft lip, cleft palate, and club foot are thought to be multifactoral. Esophageal atresia is thought to arise from a random, generalized insult to embryogenesis. Most cases of polydactyly or syndactyly have a genetic cause. The wide range of birth defects for which statistically significant excesses were noted argues against a single causative physical or chemical agent.

VI.C.2 Low Birthweight

To determine if there was an excess number of low birthweight babies born in each of the four counties (Armstrong, Carson, Potter, and Randall), the number low birthweight babies born in each county with the number that would be expected based on State rates was compared. Statistically significant excesses in the number of low birthweight babies born in Carson and Potter counties for the time period 1990 to 1994 (Appendix E; Table 8) was found. Based on the number of observed and expected cases and the narrow confidence intervals, we are confident in the precision of these results. However, the slight but statistically significant excesses are difficult to interpret; particularly because gestational age could not be accounted for in this analysis and local reporting practices may partially be responsible for these findings.

Discussion

The majority of births in this area are delivered at Northwest Texas Hospital (NTH). Evidence indicates that NTH strictly follows Texas Department of Health policy to file a birth certificate for all live births, regardless of gestational age.(15) Discussions with staff at NTH indicated that there was a strict and well-established practice for delivery room staff to send all infants with any life signs, regardless of gestational age, to neonatology. Because infants sent to neonatology would be considered live births, birth certificates would be filed. Some hospitals may report very premature infants as fetal deaths or may not report premature births with gestational ages less than 20 weeks. Thus, the practice of reporting very premature infants as live births may not be uniform throughout the State. Birth records from 1991-1994 documented a higher rate of births of infants less than 24 weeks gestation at NTH (4.85 per 1,000 live births) compared to the State of Texas (1.96 per 1,000 live births).(15) Because gestational age could not be accounted for in this analysis we cannot rule out differences in gestational age as a potential explanation for these results.

Although reporting practices may not completely explain the elevated rates for this area, there are several risk factors that increase a woman's chances of giving birth to a baby born with a low birthweight. Mothers who smoke, drink alcohol, abuse drugs, or have poor nutritional habits during pregnancy may have an increased risk for low birthweight babies. In addition, maternal age and the number of previous live births, are other examples of factors that can effect the developing fetus. In summary, there are many factors some of which can be controlled by the mother and others which cannot that play an important role in the health of the developing fetus.

VI.C.3 Cancer

To determine if there is an excess of specific types of cancer in the counties of interest, the observed number of cancer cases (incidence) and cancer deaths (mortality) for each selected type of cancer to what would be "expected" were compared. For cancer incidence, the expected number of cases was determined using race-, sex-, and age-specific cancer incidence rates for California during 1988 to 1992. California rates were used because cancer incidence data for Texas are incomplete for that time period. Bone cancer rates were not available for California, therefore, cancer incidence rates for five regions of Texas during 1985 to 1991 were used to calculate expected bone cancer cases. For cancer mortality, the expected number of deaths was determined using the mortality experience of the State of Texas.

Cancer Incidence

No significant excess of cancer incidence was observed among male or female residents of Armstrong or Carson counties (Appendix E; Tables 9 and 10) during the period 1985-1993. (16) However, a statistically significant excess of cancer incidence was observed for all cancer sites combined among females in the combined counties of Potter and Randall (Appendix E; Table 11) during the same time period. The public health significance of this excess is difficult to evaluate. The calculated Standardized Incidence Rate (SIR) is based on a large number of observed and expected cases (4,015 and 3,693 cases; respectively). Because the 95% confidence interval is narrow, we are fairly confident in the precision of this estimate of the SIR; however, the SIR barely exceeds one (1.0). Because the SIR barely exceeds one and the category "all sites" is nonspecific and includes all types of cancer; we would interpret this result cautiously. The statistically significant excess could be the result of slight (nonsignificant) elevations among many different cancer sites or the result of a large increase in one or two sites. Interpreting increases in all cancer sites combined is further complicated by the fact that known causes for individual cancer sites differ.

In a previous evaluation of cancer incidence data for this area, an analysis of the four major cell-specific types of leukemia [acute lymphocytic leukemia (ALL), chronic lymphocytic leukemia (CLL), acute myelocytic leukemia (AML), and chronic myelocytic leukemia (CML)], showed a significant excess of CLL during the period 1986-1992 among females in the combined counties of Potter and Randall. To maintain consistency in the investigation of cancer data in this area, the major leukemia subgroups were analyzed again for the period 1985-1993. There was a significant excess of CLL among males in the combined counties of Potter and Randall; however, unlike the previous investigation, the number of CLL cases among the female residents in these two counties was within the expected range (Appendix E; Tables 12 and 13).

CLL appears to have a strong inherited susceptibility component and is primarily confined to the late middle-aged and elderly (the majority of these cases were in individuals age 65 years or older). It is very rare in Asians and remains rare in Asians who migrate to western nations, which indicates the importance of genetic factors. Unlike other forms of leukemia, CLL has not been causally related to radiation exposure. The significant excess of CLL among males is a new finding as was the excess of CLL among females in the 1994 study and deserves further investigation.(11)

In response to a resident's request, the TCR analyzed the incidence of one additional cancer (malignant pleural mesothelioma). Mesothelioma is a rare form of cancer with about 2,000 cases diagnosed in the United States every year. Most people who develop this form of cancer have a history of exposure to asbestos. For the period of 1985-1993, 16 cases of mesothelioma were diagnosed in the four counties of concern (Armstrong: 0, Carson: 1, Potter and Randall: 15). Based on the U.S. age-adjusted rate, the expected rate is one per 100,000 or 14.8 cases for the four county area. The 16 observed cases are well within the range that would be expected for the area (SIR = 1.08, 95% CI = 0.62, 1.76).

Cancer Mortality

For cancer mortality, no significant excess was observed among male or female residents of Armstrong or Carson counties (Appendix E; Tables 14 and 15) during the period 1985-1994. A statistically significant excess of deaths from prostate cancer was observed among males in Potter and Randall counties (Appendix E; Tables 16 and 17). Although we cannot directly explain this excess, it is important to note that it is not consistently supported by incidence data. The differences seen in prostate cancer incidence and mortality could result from past under-reporting of prostate cancer cases (prostate cancers are now more readily diagnosed in non-hospital settings), or could be explained by differences in early diagnosis and treatment in this area compared to the rest of the State.

A statistically significant excess number of deaths due to all cancer sites was observed for males in Potter County. The public health significance of this excess is difficult to evaluate. The calculated SMR is based on a large number of observed and expected deaths (1,140 and 1,045 deaths; respectively). Because the 95% confidence interval is narrow, we are fairly confident in the precision of this estimate of the SMR; however, the SMR barely exceeds unity (1). Because the SMR barely exceeds one and the category "all sites" is nonspecific and includes all types of cancer; we would interpret this result cautiously. The statistically significant excess could be the result of slight (nonsignificant) elevations in deaths from many different types of cancer or the result of a large increase in deaths from one or two types of cancer. Interpreting increases in deaths from all types of cancer combined is further complicated by the fact that known causes for individual types of cancer differ.

Discussion

The TCR found a statistically significant excess of cancer incidence for all cancer sites combined among females in the combined Potter County/Randall County area. When such increases are observed it is impossible to determine whether the increase is the result of slight elevations among individual cancer sites or the result of a large excess in one or two cancer sites which results in a significant overall increase. Based on available information, it would be difficult to provide, with any degree of certainty, a meaningful interpretation for these results.

There was a significant excess of CLL among males in the combined counties of Potter and Randall. CLL is a specific type of cancer. Considering the previous finding of an excess of CLL among female residents from these two counties, we believe that the occurrence of CLL in these counties deserves further investigation.

For cancer mortality, a statistically significant excess of deaths from prostate cancer was observed among males in Potter and Randall counties. This finding was not consistent with the incidence data. The differences seen in prostate cancer incidence and mortality could result from past under reporting of incident prostate cancer cases or could be explained by differences in early diagnosis and treatment in this area compared to the rest of the State.

Cancer is a very common disease, much more common than most people realize. Approximately four out of every ten persons alive today will be diagnosed with some type of cancer in their lifetime. Furthermore, cancer is not one disease, but many different diseases. Different types of cancer are generally thought to have different causes. In Texas, as in the United States, cancer is the second leading cause of death, exceeded only by heart disease. In 1995, 31,571 Texans died of cancer. Sixty-five percent of these deaths were in persons 65 years of age or older. Finally, it takes time for cancer to develop, usually 20 to 40 years. Conditions that have prevailed for only the last 5 or 10 years are unlikely to be related to the current incidence of cancer in a community.

The occurrence of cancer may vary by race/ethnicity, gender, the type of cancer, geographic distribution, population under study, and a variety of other factors. Scientific studies have identified a number of factors for various cancers which may increase an individual's risk of developing a specific type of cancer. Some of these factors are: heredity, diet, age (cancer risk increases with age), family history, exposure to certain chemicals (only a limited number of chemicals show definite evidence of human carcinogenicity, e.g., benzene, asbestos, vinyl chloride, arsenic, aflatoxin), radiation (ionizing radiation and ultraviolet radiation), alcohol, and tobacco smoke.

The chances of a person developing cancer as a result of exposure to an environmental contaminant are actually slight. Tobacco use and diet are estimated to cause 30% and 35% of all cancer deaths respectively. Pollution and occupational exposures are estimated to collectively cause 4-6% of all cancer deaths.

VI.C.4 Muscular Dystrophy, Multiple Sclerosis, Amyotrophic Lateral Sclerosis, and Lupus Erythematosus

Mortality due to muscular dystrophy, multiple sclerosis, amyotrophic lateral sclerosis, and lupus erythematosus for Armstrong, Carson, Potter, Randall, and Potter/Randall counties combined for the time period of 1980-1994 are summarized in Appendix E; Tables 18-22. To determine if there was an excess of deaths due to each of these diseases, the number of deaths observed in this area for each disease was compared to the number of deaths that would be "expected" based on the age-, sex-, and race/ethnicity-specific mortality experience for the State of Texas. Standardized mortality ratios (SMRs) were calculated and evaluated as described previously in this report. SMRs were not calculated when no cases of the disease were reported.

In both Carson and Potter counties there was a statistically significant elevation in the number of observed deaths due to muscular dystrophy for males (Appendix E; Table 19-20). In Randall County and in Potter/Randall counties combined, there was a statistically significant elevation in the number of observed deaths due to multiple sclerosis for females (Appendix E; Table 21-22). For males in Potter/Randall counties combined, there was a statistically significant excess of amyotrophic lateral sclerosis. Deaths due to lupus erythematosus were not found to be excessive in any of the four counties of concern.

Discussion

Incidence data were not available for these diseases therefore we used mortality data. In general, mortality data for these diseases are difficult to interpret because they are diseases that people may have for many years with death often resulting from other causes. Caution must be exercised when interpreting the elevations among the four county area. All of the counties are sparsely populated and even one additional death can result in an elevated mortality ratio. The causes associated with each of these diseases are still very speculative. Because of the nature of these data we are not able to provide, with any degree of certainty, a meaningful interpretation.

Genetics is the primary risk factor associated with muscular dystrophy. This means that individuals who have relatives who had the disease, may have a greater chance of developing the disease themselves.

Multiple sclerosis, like muscular dystrophy, also may have a genetic component associated with its development. In addition to the genetic component, some other factors that have potentially been associated with this disease include: exposure to viral agents, country of origin, exposure to heavy metals, and being female.

Several possible risk factors that may be associated with amyotrophic lateral sclerosis have been discussed in the scientific and medical literature. In addition to genetics, prior trauma, heavy milk consumption, heavy physical labor, occupational exposure to chemicals involved in the manufacture of plastics, and exposure to infectious agents prior to the onset of disease have all been implicated in the development of amyotrophic lateral sclerosis.

Lupus erythematosus, like the other diseases, also has a strong genetic association with its development. In addition, exposure to specific chemicals and toxins, dietary factors, and exposure to infectious agents prior to the onset of disease have been implicated in the development of lupus.

VI.D ATSDR Child Health Initiative

The Agency for Toxic Substances and Disease Registry's (ATSDR) Child Health Initiative recognizes that the unique vulnerabilities of infants and children demand special emphasis in communities faced with contamination of their water, soil, air, or food. Children are at greater risk than adults from certain kinds of exposures to hazardous substances emitted from waste sites and emergency events. They are more likely to be exposed because they play outdoors and they often bring food into contaminated areas. They are shorter than adults, which means they breathe dust, soil, and heavy vapors close to the ground. Children are also smaller, resulting in higher doses of chemical exposure per body weight. The developing body systems of children can sustain permanent damage if toxic exposures occur during critical growth stages. Most importantly, children depend completely on adults for risk identification and management decisions, housing decisions, and access to medical care.

ATSDR evaluated the likelihood for children, living within 5 miles of the Pantex Plant, to be exposed to plant contaminants at levels known to cause adverse health effects. ATSDR did not identify any current exposure pathways for children. The perimeter of the plant is fenced and guarded, preventing children from wandering into areas where environmental contamination exists. Unless these conditions change, there should not be any future exposure situations in which children may be exposed to chemical contaminants from the Pantex Plant.

The Texas Department of Health (TDH) evaluated health outcome data for birth defects, low birth weight, and specific cancers of community concern. These evaluations are discussed in the Health Outcome Data and Health Outcome Data Evaluation sections. Though there are no pathways identified which may expose children, TDH recommended surveillance of birth defects at area hospitals and birthing centers to better compare rates to other surveillance systems in the United States.

Using information obtained from birth certificates, fetal death certificates, and infant death certificates, the Texas Department of Health found that the observed number of cases and/or deaths for several categories of birth defects in this area are higher than what would be expected based on comparison to State rates. Although better reporting practices could not be ruled out as a possible explanation for the higher than expected number of cases in this area, TDH also could not rule out the possibility that the apparent excesses are real. TDH is currently in the process of expanding an active birth defect surveillance system into this area beginning with 1998 deliveries. Data from this active surveillance system will be used to evaluate the accuracy and completeness of vital records in this area.

Information obtained from birth certificates indicates the rate of low birth weight babies born in Carson and Potter counties was higher than the rate reported for the State of Texas. Northwest Texas Hospital, the main birthing center for this area, strictly adheres to TDH policy to file a birth certificate for all live births regardless of gestational age. Previous discussions with hospital staff indicate that all infants with any life signs, regardless of gestational age, are considered live births. This practice of reporting very premature births as live births may not be uniform throughout the State. Some hospitals may report very premature births as fetal deaths, and some may not report premature births with gestational ages less than 20 weeks.

Next Section          Table of Contents

Agency for Toxic Substances and Disease Registry, 1825 Century Blvd, Atlanta, GA 30345
Contact CDC: 800-232-4636 / TTY: 888-232-6348 

Department of Health and Human Services