Exposure to VOCs at Superfund Sites Results in High Medical Costs 

Investigators at the Agency for Toxic Substances and Disease Registry (ATSDR) and the Joint Institute for Energy and Environment (JIEE) recently conducted an analysis to estimate some of the medical and lost productivity costs from health conditions occurring in communities located near Superfund sites that are contaminated with volatile organic compounds (VOCs). The results were published in the environmental medicine journal Environmental Research (1). The analysis estimated the general magnitude of the costs, rather than a precise amount. Costs for 258 Superfund sites were calculated to be approximately $300 million per year. This figure does not include the costs of any remediation or public health intervention programs.

The sites chosen had evidence of a completed human exposure pathway. The analysis was conducted in four steps:
(1)  Health conditions - Identify health conditions that have been found to be occurring in excess in people who have known exposure to VOCs in drinking water or who live near a contaminated site.
(2)  Population estimate - Estimate the total population living within one-half mile of the 258 Superfund sites.
(3) Costs of each health condition - Estimate the average medical costs and average costs of lost productivity for each incident of an adverse health effect or illness. 
(4)  Economic burden - Calculate the total dollar amount of the health burden for all sites.

Health Conditions
The health conditions found to be associated with a known exposure to VOCs in drinking water or with living near contaminated sites include several types of birth defects, stroke, and several general illnesses reported in excess among persons registered in ATSDR’s National Exposure Registry (NER). Recent studies have identified associations between VOC exposures and birth defects. One study done in New Jersey by Bove et al. (2) reported that maternal exposure to VOC-contaminated drinking water systems was associated with increased risks of neural tube defects, oral cleft defects, and major cardiac malformations. These conditions were found to be associated with the exposure parameter (odds ratio of 1.5). These findings are consistent with other reports. 

The Bove et al. paper reported an odds ratio of 2.1 for benzene and 2.5 for trichloroethylene (TCE). The ATSDR-JIEE study uses the lower of the two odds ratios found in the Bove et al. article. The primary source of data for the estimates of health conditions other than birth defects was ATSDR’s NER. Substance-specific categories exist in the NER for three VOCs (i.e., trichloroethylene, trichloroethane [TCA], and benzene). 

Population Estimate
National Priorities List (NPL) sites were chosen for this analysis based on environmental testing, which confirmed, at minimum, the presence of TCE, TCA, or benzene in a completed human exposure pathway, and ATSDR’s public health assessment that concluded that the site poses a public health hazard. A geographic information system (GIS) analysis was used to estimate the population at each of the 258 sites (3). Spatial analysis was used to calculate the population located within one-half mile of the site boundary for each site. Data were available to define boundaries for 225 of the sites. The statistical distribution of the population estimates of these 225 sites was skewed. To identify a representative population size for these sites, the sites were divided into nine groupings of 25 sites each. The median or fifth grouping had a mean population estimate of 1,600 people per site. This value was used as an estimate of the population for the 32 sites for which there were no data on the boundaries. The estimated population for all 258 sites totaled approximately 1.72 million residents.

The live birth rate was estimated to be 0.0167 per person. The percentage of children younger than 10 years of age was estimated to be approximately 14.6% of the U.S. population based on the 1990 age-specific population estimates from the National Center for Health Statistics (4). For the 258 sites, it was estimated that approximately 28,700 live births occurred each year and that there were 251,000 children younger than 10 years of age. 

Costs of Each Health Condition
Medical costs and the costs of lost productivity were estimated for birth defects, strokes, urinary tract disorders, diabetes, eczema and skin conditions, anemia, and speech and hearing impairments. 

Birth Defects. For birth defects, a study by Waitzman et al. was used for data on the average costs of the more costly birth defects (5). The estimated costs of those birth defects not covered by Waitzman et al. were based on Waitzman et al. data for the lowest cost conditions covered within each birth defect category. Average sex-specific earnings for each age were taken into account for lost productivity estimates. Costs were adjusted from the year 1988 (used by Waitzman et al.) to 1995 using the Consumer Price Index component for medical services and the nonfarm private wage inflator (6, 7). Future cost estimates were summed into an estimate of their total present value and discounted at 5% annually, based on the recommendation by the prevention effectiveness program of the Centers for Disease Control and Prevention (8).

Lifetime costs for birth defects based on an estimated 210 excess cases in a given year were estimated to be approximately $68 million. 

Strokes. Because data were not available on the proportion of people who had a stroke who incurred medical and related costs in any given year, information in ATSDR’s TCE Subregistry was used to estimate this proportion. The cost estimate for stroke was based on excess prevalence. Calculations were based on the proportion of stroke victims who had a first stroke and who received hospital care within the 12 months prior to TCE Subregistry data collection. This proportion was used as a surrogate estimate. It underestimates the actual proportion because it omits those whose first stroke occurred prior to the 12 months preceding the date when the subregistry data were collected. Also, no data were available on the actual outpatient costs for the same timeframe. Data from MarketScan (9) and the 1980 National Survey of Stroke (10) were  used to make these estimates. MarketScan data were used to calculate inpatient and outpatient costs and lost productivity because of hospitalization and outpatient care (9). National Survey of Stroke data suggested that long-term outpatient care is as expensive as inpatient care and that costs for stroke-related morbidity were about 15% of inpatient costs. Mortality-related costs also were not available, leading to a significant underestimation of total costs, even after lost productivity is taken into account.

Based on an estimated 8,600 excess cases of stroke each year, total annual cost for these cases is $60 million. Taking into account additional factors not formally considered in overall costs, such as long-term care, lost earnings, and lost productivity, total cost for stroke victims could reach $120 million. 

Other Health Conditions. Urinary tract disorders, diabetes, eczema and skin conditions, anemia, and speech and hearing impairment were also considered to be in excess, based on TCE Subregistry data. MarketScan data from 1992 were used to estimate the average inpatient, outpatient, long-term care, lost productivity, and other costs that a person with one of these conditions would have in one year. MarketScan data were used for the annual costs of inpatient and outpatient care for each health condition; information on approximately 2.9 million people was available from the database (9). Data from the American Diabetes Association (ADA) were used to calculate costs of secondary conditions, long-term care, and lost productivity (11). Costs from lost earnings or productivity from premature mortality were calculated for diabetes and stroke (10, 11).

Total estimated medical costs and lost work time costs totaled $39 million for one year for health conditions other than stroke and birth defects. For example, the costs of urinary tract disorders  amounted to an estimated $14 million for medical expenses and lost work time because of inpatient or outpatient care. Anemia ($9.6 million) and diabetes ($7.4 million) were the next most expensive conditions. Additional costs estimated for diabetes, including complications and associated comorbid conditions, long-term care not covered by the primary insurer, and additional lost productivity from morbidity and premature mortality, could amount to $43 million. 

Economic Burden
The total estimated cost of $330 million represents the medical, long-term care, and lost-productivity costs resulting from these excess numbers of selected health conditions at the chosen 258 VOC-contaminated NPL sites. (See Table 1 for overview of costs.) Of this figure, $170 million covered medical costs, long-term care costs for birth defects, and lost productivity due to time lost due to inpatient and outpatient care. These costs are attributable to the excess incidence rates for birth defects and to the excess prevalence of other health conditions. Excess rates of birth defects ranged from 0.9 per 1,000 live births for neural tube defects to 4.7 per 1,000 live births for cardiac malformations. Estimates of the excess numbers of cases for illnesses include 5 per 1,000 people for diabetes, 26 per 1,000 children younger than 10 years of age for speech impairments, and 5 per 1,000 for stroke.

^*  Excess cases are expressed as the number of newborns with birth defects, per year, above the number expected.
^†  Excess cases expressed as the number of cases in a given year, above the number expected.
^‡  If long-term care costs, lost earnings from premature mortality, and lost productivity from morbidity are not included, then the total costs of the excess cases of stroke could be $60 million.
^§  Includes urinary tract disorders, anemia, diabetes, eczema and skin diseases, and speech and hearing impairment.
^£   If long-term care costs, lost earnings from premature mortality, and lost productivity from morbidity are not included, then the total costs of the excess cases of other health conditions could be $39 million.

Discussion
The results are indicative, but not definitive. Uncertainty exists because

• there is no definitive proof of a causal relationship between illness and exposure to VOCs;
• illness rates for conditions other than birth defects were based upon self-report questionnaires;
• the exposed population numbers used are estimates; 
• the costs are representative, not exact; and
• the total economic burden of these cases is probably underestimated because of other types of costs not considered (e.g., quality of life and decreased property values near sites). 

For more information, please contact Jeff Lybarger, Director, Division of Health Studies, Agency for Toxic Substances and Disease Registry (e-mail jal2@cdc.gov) at  404-639-6200.

References

1. Lybarger JA, Lee R, Vogt DP, Perhac Jr RM, Spengler RF, Brown DR. Medical costs and lost productivity from health conditions at volatile organic compound-contaminated Superfund sites. Environ Res 1998;79:9–19.
2. Bove FJ, Fulcomer MC, Klotz JB, Esmart J, Dufficy EM, Savrin JE. Public drinking water contamination and birth outcomes. Am J Epidemiol 1995;141:850–62.
3. ARC/INFO. Version 6. Redlands (CA): Environmental Systems Research Institute;1991.
4. National Center for Health Statistics. Health in the United States 1993. Hyattsville (MD): U.S. Department of Health and Human Services, Public  Health Service, Centers for Disease Control and Prevention;1994.
5. Waitzman N, Romano PS, Scheffler RM. Estimates of  the economic costs of birth defects. Inquiry 1994;33:188–205.
6. Council of Economic Advisors. Consumer price index component for medical services. Economic indicators. Washington (DC): U.S. Government Printing Office; 1995.
7. Council of Economic Advisors. Non-farm private wage deflator. Economic indicators. Washington (DC): U.S. Government Printing Office; 1995.
8. Centers for Disease Control and Prevention. A practical guide to prevention effectiveness: decision and economic analyses. Atlanta (GA): U.S.  Department of Health and Human Services, Public Health Service; 1994.
9. MarketScan. The MEDSTAT group, all rights reserved; 1995.
10. Westat. National survey of stroke. U.S. Department of  Health, Education, and Welfare Publication. Washington (DC): National Institutes of Health 80– 2069;1980.
11. American Diabetes Association. Direct and indirect costs of diabetes in the United States in 1992. Alexandria (VA): American Diabetes Association; 1998.New York Department of Environmental Conservation. Ambient air quality characterization of Fresh Kills Landfill, 1994 preliminary report. New York: New York Department of Environmental Conservation; 1995 Feb. 

Studies Suggest Link Between Birth Defects and Maternal Residence Near Waste Sites 

Does living near a hazardous waste site increase the risk for birth defects? Over the past decade, six studies have been conducted to investigate this question. All of these studies used population-based birth defects registries to confirm the cases. Four of these studies were funded by the Agency for Toxic Substances and Disease Registry (ATSDR). 

The studies did not provide a clear and definitive answer to this question, primarily because of methodologic limitations. Summarizing the findings from these studies is also difficult because they did not always evaluate the same birth defect groupings. Nevertheless, as detailed below, the studies provided reasonable evidence to establish an association between maternal residential proximity to hazardous waste sites and increased risks for particular birth defects diagnosed within the first year of life.

An ATSDR-supported study at Times Beach, Missouri (1) evaluated the effect of living near areas with dioxin-contaminated soil. A total of 402 infants born during 1972–1982 to mothers residing in a contaminated area were compared with 804 infants born to mothers residing in other areas of the state, which had not dioxin contamination. The groups were matched on mother’s age and race/ethnicity of the infants.

Of the potentially dioxin-exposed (Times Beach) infants, three had central nervous system (CNS) defects, constituting a threefold increased risk compared with unexposed infants. Four infants residing in Times Beach had gastrointestinal defects and three had undescended testicular defects, constituting a twofold increased risk for each defect. Although the association between maternal residential proximity to dioxin-contaminated soil and birth defects was obviously weak because of the small number of infants included in the study, the results should not be dismissed. In addition, the finding of an increased risk for CNS defects agrees with findings from studies of children born to parents exposed to dioxin-contaminated Agent Orange during the Vietnam War (2). The five other studies (two each conducted in New York and California, and one in Europe) focused on numerous toxic waste sites over a wide geographic area. A study conducted in New York evaluated data from 590 sites in 20 counties, excluding New York City and rural counties (3). A total of 9,313 infants with registered defects and a random sample of 17,802 infants without registered defects born during 1983–1984 were included in the study. Maternal residence and relevant toxic waste sites were mapped by latitude and longitude to measure distances between them. Residence within 1 mile of a site was associated with a 29% increased risk for CNS defects, a 16% increased risk for musculoskeletal defects, and a 32% increased risk for skin problems. 

The exposure potential (high, low, or none) of each site was determined based on scores derived from a modified version of the EPA’s Hazard Ranking System (HRS) which is used to prioritize sites for remediation. Residence within 1 mile of a site classified as having a high exposure potential was associated with a 48% increased risk for CNS defects, a 75% increased risk for musculoskeletal defects, and a more than twofold increased risk for skin problems. Potential exposure to metals and solvents from the sites was associated with increased risks for CNS defects of 34% and 24%, respectively. Potential exposures to pesticides from the sites increased the risk for musculoskeletal defects by 20%. 

A follow-up study funded by ATSDR focused on 473 CNS and 3,304 musculoskeletal defects registered in 18 of the original 20 counties during 1983–1986 (4, 5). Exposure assessment was enhanced in this study by dividing the 1-mile radius around each site into 25 sectors and determining each sector’s exposure potential from four pathways (air vapor from soils and surface waters, air particulates, private wells, and air vapors via basements) based on available sample data and other information (e.g., prevailing wind data from the nearest weather station). Three chemical classes were evaluated: solvents, pesticides, and metals.

Similar to the earlier NY study, residence near a toxic waste site (within 0.6 miles) was associated with slightly increased risks for CNS defects (18%) and for musculoskeletal defects (11%). Potential exposure to metals from the sites was linked to a 25% increased risk for CNS defects. Unlike the earlier study, no association was found between musculoskeletal defects and potential exposure to pesticides or between CNS defects and potential exposure to solvents.

One limitation of the NY studies was that an evaluation of the CNS subgrouping, neural tube defects (NTDs), was not performed. A second limitation was that their assessments of exposure potential at each site were based on sampling data of widely differing quality that were collected for other purposes (i.e., remediation and enforcement). For some sites, there were no available sample data to assess exposure potential. Moreover, assessing the exposure potential of several hundred sites could only result in superficial determinations. These two New York studies suggested that maternal residential proximity to toxic waste sites was associated with small increases in risk for CNS and musculoskeletal birth defects.

The two California-based studies were also funded by ATSDR. The first study compared 5,046 infants with registered birth defects with a random sample of 28,085 infants without registered birth defects (6). The infants were born during 1983–1985 in the five-county San Francisco Bay area, which contains 300 toxic waste sites. Maternal residence and toxic waste sites were mapped to census tracts, and the exposure potential of each site was based on whether there was evidence of off-site contamination and/or direct public contact with the site.

Maternal residence in a census tract containing a site with an exposure potential was not associated with all CNS defects; however, an almost twofold increased risk was found for the CNS defect spina bifida, which is a NTD. An increased risk of 50% was found for all heart defects, including conotruncal defects. When exposure pathways (e.g., air, drinking water) from the sites were evaluated, higher risks were found in general, but these findings should be considered suggestive because the evaluation was at the census tract level and might not reflect accurately the exposure risks to individuals residing in proximity to the sites.

The second California study covered most of the state and focused on NTDs, oral cleft defects, and conotruncal heart defects (7). The birth defect cases, control birth cases, and 105 National Priorities List (NPL) sites were mapped to latitude and longitude coordinates. 

The study had two parts. The first part included 538 births with registered NTDs and 537 control births occurring between 1989 and 1991. The second part included 664 births with registered birth defects (458 oral cleft defects, 206 conotruncal defects) and 481 control births occurring between 1987 and 1988. Interviews were conducted to determine mother’s first trimester residence and to obtain information on a wide range of risk factors such as smoking and occupational exposures. 

The NPL sites were characterized by exposure potential (yes/no) and by the media contaminated (e.g., air, soil, groundwater). The study found that first trimester residence within a quarter mile of an NPL site with exposure potential was associated with a twofold increased risk for NTDs, a fourfold increased risk for conotruncal defects, and a 20% increased risk for oral cleft defects. As the distance from the NPL sites increased from ¼ to 1 mile, the risks declined substantially for NTDs and conotruncal defects. There was no relationship between distance from the sites and the risk of oral cleft defects.

The assessment of the exposure potential at each site suffered from the same limitations as occurred in the NY studies. Nevertheless, this study provided reasonable evidence that the risks for NTDs and conotruncal heart defects increased substantially in residential areas bordering toxic waste sites.

A recent study conducted in five European countries evaluated births within 7 kilometers (km) of 21 toxic waste sites over the period when the regional birth defect registries in each country started operation until December 1993. In Italy and Denmark, the study period began in the mid-1980s, and for the United Kingdom, Belgium, and France, 1990. A total of 1,089 birth defect cases and a matched sample of 2,366 control births without birth defects were included in the study (8). Births to mothers residing at the time of birth within 3 km of a site were considered exposed to the site. No attempt was made to determine each site’s exposure potential.

Residence within 3 km of a site was associated with a 86% increased risk for NTDs, increased risks for some of the major heart defects, a 63% increased risk for cleft palate, an almost twofold increased risk for skin problems and for hypospadias, a more than twofold increased risk of anomalies of the esophagus (e.g., tracheoesophageal fistula), and a more than threefold increased risk of an abdominal wall defect (gastroschisis). Slightly increased risks were also found for cleft lip and limb reduction defects. A major limitation of this study was the crude manner in which exposure to each site was determined (i.e., residing within 3 km of a site at time of birth). In addition, heart defects were grouped differently in this study so that the results were not directly comparable to other studies. Nevertheless, the increased risks found for NTDs and several major heart defects were consistent with the findings in the California studies and the Times Beach study.

To strengthen the evidence for the associations between toxic waste site exposures and increased risks for specific birth defects, future studies must determine more precisely the exposures resulting from toxic waste sites. One approach is to include in a study only those sites that have adequate off-site sample data and that are highly likely to cause exposures to surrounding populations. Sites with similar exposure pathways and/or contaminants could be grouped together. Although this approach may severely restrict the number of sites evaluated, and thereby reduce the study’s statistical power, it would also reduce considerably the amount of exposure misclassification bias in the study. Another approach is to focus on a known complete exposure pathway, e.g., public drinking water contaminated by pollutants from industrial facilities, agricultural runoff, and waste sites.

For more information, contact Frank Bove, ScD, ATSDR, 1600 Clifton Rd, NE, MS E31, Atlanta, GA 30333; telephone (404) 639-5126; fax (404) 639-6219; e-mail fjb0@cdc.gov.

References
1.  Stockbauer JW, Hoffman RE, Schramm WF, Edmonds LD. Reproductive outcomes of mothers with potential exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Am J Epidemiol 1988;128:410–19.
2. Institute of Medicine. Veterans and Agent Orange: Update 1998. Washington, DC: National Academy Press, 1999.
3. Geschwind SA, Stolwijk JAJ, Bracken M, et al. Risk of congenital malformations associated with proximity to hazardous waste sites. Am J Epidemiol 1992;135:1197–207.
4. Marshall EG, Gensburg LJ, Geary NS, et al. Analytic study to evaluate associations between hazardous waste sites and birth defects. New York, NY: New York State Department of Health and Health Research, Inc., Bureau of Environmental and Occupational Epidemiology; June 1995.
5. Marshall EG, Gensburg LJ, Deres DA, et al. Maternal residential exposure to hazardous wastes and risk of central nervous system and musculoskeletal birth defects. Arch Environ Health 1997;52:416–25.
6. Shaw GM, Schulman J, Frisch JD, et al. Congenital malformations and birthweight in areas with potential environmental contamination. Arch Environ Health 1992;47:147–54.
7. Croen LA, Shaw GM, Sanbonmatsu L, et al. Maternal residential proximity to hazardous waste sites and risk for selected congenital malformations. Epidemiol 1997;8:347–54.
8. Dolk H, Vrijheid M, Armstrong B, et al. Risk of congenital anomalies near hazardous-waste landfill sites in Europe: the EUROHAZCON study. Lancet 1998;352:423–27.

[Table of Contents]

Children and the Environment

Principal staff include Scott Barnhardt, MD, MPH, and Sanders Chai, MD, both of the Occupational and Environmental Medicine Program, University of Washington Harborview Medical Center; and William Robertson, MD, medical director of the Washington Poison Center and professor of pediatrics at the University of Washington.

The Occupational and Environmental Health Center at Cambridge Hospital collaborates with the Occupational and Environmental Health Center, the Pediatric Environmental Health Center at Children’s Hospital in Boston, the Center for Occupational and Environmental Medicine of the Massachusetts Respiratory Hospital, and the Occupational and Environmental Medicine Program at the Harvard School of Public Health. 

Principal staff include Rose Goldman, MD, MPH, director of  Occupational and Environmental Health Program at  Cambridge Hospital; Howard Hu, MD, professor of Occupational and Environmental Medicine at Harvard School of Public Health; Michael Shannon, MD, MPH, associate chief of the Division of Emergency Medicine and director of the Lead Clinic of Children’s Hospital; and Alan Woolf, MD, MPH, associate in medicine (general pediatrics) and director of the Clinical Toxicology Program at Children’s Hospital.

Mount Sinai Pediatric Environmental Health Center is located at the Mount Sinai-Irving J. Selikoff Center for Occupational and Environmental Medicine and works with ATSDR and the National Center for Environmental Health.

Principal staff include Joel Forman, MD, Department of Pediatrics, Mt. Sinai Medical Center; Phillip Landrigan, MD, MPH, chairman of the Department of Community and Preventive Medicine at the Mt. Sinai Medical Center and director of Occupational and Environmental Medicine. 

Cook County Pediatric Environmental Health Specialty Unit has joined efforts with Cook County Hospital, Toxikon Consortium (a toxicology consortium that links the toxicology programs of Cook County Hospital, Rush-Presbyterian-St. Luke’s Medical Center, the University of Illinois, and the Illinois Poison Center), Great Lakes Center of Occupational and Environmental Safety and Health, and University of Illinois at Chicago School of Public Health. 

Principal staff include Carl Baum, MD, Department of  Pediatrics at Children’s Memorial Hospital; Daniel Hryhorczuk, MD, MPH, director of the Great Lakes Center of Occupational and Environmental Safety and Health, director of the Toxikon Consortium, and chief of the Section of Clinical Toxicology at Cook County Hospital; and Myrtis Sullivan, MD, Department of Pediatrics at Cook County Hospital, formerly director of the Cook County Hospital Pediatric Emergency Room.

[Table of Contents]

New Pediatric Case Study in Environmental Medicine

• understand how and why children differ from adults in susceptibilities to environmental hazards,
• incorporate knowledge of environmental medicine in their evaluation of well and sick children,
• remember to ask parents about their occupation as part of a child’s environmental history,
• access additional resources on environmental medicine and advice, and 
• learn about different exposure sources for children.