SECTION 1
INTRODUCTION
This is a report on the Baseline, Followup 1, and Followup 2 activities and findings from the analyses of data collected from participants in the Dioxin Subregistry. The Dioxin Subregistry is part of the National Exposure Registry, which was created and is being maintained by the Agency for Toxic Substances and Disease Registry (ATSDR).
In 1988, the policies and procedures proposed for the National Exposure Registry were reviewed extensively by several committees of independent scientists, state representatives, representatives of other federal agencies, and other interested people. The policies and procedures were published in the National Exposure Registry Policies and Procedures Manual (1) and were revised in 1994 (2). The Dioxin Subregistry was one of the first subregistries to be established as part of the National Exposure Registry program. There are currently three chemical subregistries (trichloroethylene, benzene, and dioxin). Two additional subregistries (chromium and radioactive substances) are planned.
In keeping with Registry goals (1), this report represents the first of a set of ongoing reports and publications that will summarize the latest Dioxin Subregistry findings and suggest specific hypotheses for future research. The research will focus on these and other residential populations that have experienced similar exposures to 2,3,7,8-tetrachlorodibenzo-p-dioxin, hereafter referred to as“dioxin”. This report highlights some health outcomes and confounding variables that should be considered for analysis in future epidemiological or health studies.
The Policies and Procedures Manual (1) describes in detail all policies, procedures, and operational details pertinent to establishing the dioxin and other subregistries of the National Exposure Registry. Specific topics from the policies and procedures document are reiterated in this report, where necessary, for clarity.
Section 2 of the report provides a discussion of the information available in 1988, the year the Dioxin Subregistry was approved, on the toxicity of dioxin-the related epidemiological and toxicological studies. Section 3 presents a detailed discussion of the sites included in the Dioxin Subregistry, including information on residential sampling schemes and periods of exposure. Section 3 also provides descriptive and summary information on the environmental data and a description of the previous health studies in the Times Beach, Missouri, area. A discussion by site of the data collection period, participation rates, and number of registrants is included.
Section 4 of this report provides an overview of the characteristics and health status of registrants who took part in the Baseline effort. Section 4 also includes the comparison of the Dioxin Subregistry data with national survey data files for smoking habits and demographic characteristics and reporting rates of adverse health outcomes. Section 5 provides the same information for Followup 1 and Followup 2. Section 6 summarizes the findings of the report and discusses the findings in relationship to the published literature. Section 7 states the conclusions of the analysis of the Dioxin Subregistry Baseline data and outlines the future activities related to the subregistry.
SECTION 2
THE INCLUSION OF DIOXIN AS A PRIMARY CONTAMINANT
BACKGROUND
In the early 1970s, approximately 29 kilograms of 2,3,7,8-tetrachlorodibenzo-p-dioxin (dioxin)-contaminated sludge wastes, originating as by-products of hexachlorophene production in a southwest Missouri plant, were mixed with waste oils and sprayed for dust control throughout the state. As of 1987, almost 250 residential, recreational (including several horse arenas), and commercial areas were thought to be contaminated; 44 sites had been confirmed as having at least 1 part per billion (ppb) of dioxin in the soil (3). At one of these sites, levels as high as 33,000 ppb were measured. Isolated levels of more than 2,000 ppb were found in some contaminated areas, but most detectable levels in soil samples ranged from several hundred parts per billion to less than 1 ppb.
In August of 1971, the Centers for Disease Control (CDC) and the Missouri Department of Health (MDH) became involved in the investigation of this contamination after receiving a report of a child who had developed hemorrhagic cystitis (4) after playing in an oil-treated horse arena. In 1974, this work culminated in the laboratory identification of dioxin in the waste oil and soil. With the discovery of widespread dioxin contamination in Missouri in mid-1982, CDC, in consultation with the MDH, resumed public health activities.
By January 1983, the following public health actions had been initiated by CDC and the MDH:
1.Provision of health education for both the medical and public health communities, as well as the general public, about then current understandings of the health effects of dioxin exposure.
2.Provision of a dermatologic screening clinic to the general public.
3.Creation of a central listing of potentially exposed people.
4.Design and implementation of a pilot medical study of a "highest risk" cohort (4).
These activities led to further MDH and CDC studies, including the following:
1.Quail Run Medical Epidemiology Study (1984/5)
2.Quail Run Follow-up Study (1985/6)
3.Reproductive Outcomes Study (1985/6)
4.Adipose (Fat) Tissue Study (1986)
In 1988, the decision was made to include dioxin as a primary chemical for one of the first subregistries of the National Exposure Registry. The factors that led to the selection of dioxin were the documented presence of dioxin in the environment; published evidence of dioxin toxicity in worker populations and in toxicologic studies; the paucity of information on low-level, long-term exposures to dioxin; and the extreme national concern regarding the general health effects of dioxin. It was thought that the Registry could contribute significantly to the information available on the association of adverse health effects, should they exist, following long-term, low-level exposure to dioxin in the environment. The policies and procedures for establishing subregistries are described elsewhere (1,2).
HUMAN DATA
As of 1988, the year the Dioxin Subregistry was established, relatively little was known about the human health effects following exposure to dioxin (Figure 2-1). (Note: In this section, the discussion is limited to the information available in 1988. Information available since that time is included in Section 6). Most of the direct knowledge of the effects of dioxin exposure on human health had been obtained from the study of workers exposed to dioxin during the production or subsequent handling of 2,4,5-trichlorophenol or pesticides produced therefrom, such as 2,4,5-trichlorophenoxyacetic acid (4). Chloracne, immunotoxicity, hyperpigmentation, hyperkeratosis, hirsutism of the skin, possible hepatotoxicity, hypertriglyceridemia and hypercholesterolemia, aching muscles, loss of appetite, weight loss, digestive disorders, headaches, neuropathy, insomnia, sensory changes, and loss of libido had been observed in humans exposed to chemicals contaminated with dioxin (6). These reported effects could have been the result of dioxin exposure, or of exposure to the chemicals of which dioxin is a contaminant or to the solvents in which these compounds are normally dissolved. Because some herbicides and some industrial chemicals contain dioxin as a contaminant, the primary route of exposure was most likely dermal, although some oral and inhalation exposure also occurred.
Chloracne was the only substantiated effect produced in humans by certain compounds contaminated with dioxin. These persistent, deforming lesions of the face and upper body had been recognized for many years as resulting from exposure to certain halogenated aromatic compounds, and it was believed that dioxin was the most effective compound in producing these lesions.
Evidence suggested that dioxin was hepatotoxic in humans. In populations exposed to herbicides and other industrial chemicals contaminated with dioxin, there had been reports of increased serum levels of liver enzymes and the development of porphyria cutanea tarda (PCT), an acquired form of porphyria characterized by chronic skin lesions (6). In all studies, however, exposure could have been to chemicals that can also cause liver damage, and the diagnosis of PCT in some of the studies might have been questionable. Similarly, data suggested that dioxin might affect the immune system in humans, but the same limitations discussed with regard to hepatotoxicity applied to immunotoxicity.
Studies of human populations exposed to herbicides and other industrial chemicals contaminated with dioxin suggested that dioxin produced a variety of developmental effects. After reviewing these studies, the U.S. Environmental Protection Agency (EPA) indicated that the data were not inconsistent with dioxin adversely affecting development, but as a result of the limitations
Figure 2-1.—Availability of information on the health effects of dioxin (human data).
of the data, these studies could not prove an association with dioxin exposure and the observed effect (6). The major limitations in these human studies were the concomitant exposures to other potentially toxic chemicals, the lack of any specific quantitative data on the extent of exposure of individuals within the study group, and the lack of statistical power.
EPA also reviewed human reproductive toxicity studies of groups exposed to herbicides and other industrial chemicals contaminated with dioxin. These studies did not provide a scientifically valid indication that dioxin adversely affected either male or female reproductive performance, or that exposure to dioxin was without effect (6). The limitations of the studies are similar to those discussed previously.
In several epidemiology studies of humans exposed to herbicides contaminated with dioxin (reviewed in 6), an association was reported between exposure and soft tissue sarcomas (of various sites) and lymphomas. Although many of these studies had confounding factors, the problems with the studies were not sufficient to explain the highly significant risks of soft tissue sarcoma in workers. In addition, EPA reviewed a number of studies that were considered to be consistent with or tended to support the findings of soft tissue sarcoma in groups thought to be exposed to chemicals contaminated with dioxin. A number of these studies found no association between risk of soft tissue sarcoma and exposure to herbicides contaminated with dioxin. EPA concluded that the epidemiologic data appeared to provide limited evidence that exposure to phenoxyacetic acid herbicides or chlorophenols, or both, was causally related to the risks of soft tissue sarcoma, but none of the data were sufficient to implicate dioxin alone (6).
Information on the health effects of nonoccupational environmental exposure to dioxin and its congeners was sparse. After an explosion in 1976 at the ICMESA plant in Seveso, Italy, exposed children developed chloracne, the exposed population exhibited abnormal liver function tests, and persons with chloracne showed a statistically significantly elevated incidence of abnormal nerve conduction in tests (6). In 1971, in Missouri, a child developed hemorrhagic cystitis after playing in a riding arena containing dirt contaminated with dioxin in concentrations as high as 33,000 ppb (6).
Published reports presented the results of measurements of dioxin in human tissues. One report described the distribution of dioxin in a 55-year-old woman who was exposed to a mean soil dioxin concentration of 185.4 parts per billion dioxin for 16 days in Seveso, Italy. She died of pancreatic cancer 7 months after the explosion at the ICMESA plant, and post-mortem tissue analysis revealed dioxin levels of 1,840 parts per trillion (ppt) in fat, 150 ppt in liver, 60 ppt in brain, and 6 ppt in blood. No other tissue analyses have been reported from individuals exposed at Seveso (6).
ANIMAL DATA
The amount of animal evidence for the toxicity of dioxin (Figure 2-2), although also limited in 1988, was greater than that for humans. Four major toxic effects characteristic of dioxin are chloracne, the wasting syndrome, hepatotoxicity, and immunotoxicity. As of 1988, the latter three effects had been clearly demonstrated only in laboratory animals.
Since chloracne, the only lesion definitely identified in humans as resulting from dioxin exposure, could be detected in only a few species, the investigation of this effect had been limited. The only dermal animal data that provided quantitative information on chloracne were from a study in which hairless mice treated with dioxin developed dermal lesions that resembled some features of chloracne in humans (6).
The wasting syndrome was characterized by extreme loss of body weight. In acute oral exposure studies, this syndrome was associated with lethal doses. A dose-response relationship for the wasting syndrome had been defined in a 90-day study in guinea pigs (6).
Dioxin was hepatotoxic in all species tested; however, the severity of the lesions depended on the species studied. Although liver damage was not as severe in the guinea pig, the most sensitive species tested with regard to lethality, liver changes such as focal necrosis and hypertrophy had been observed at very low doses. There were studies of chronic exposure in guinea pigs; however, results of studies of chronic exposure in rats showed "toxic hepatitis" and degenerative changes (6).
The guinea pig also appeared to be the species most sensitive to the immunotoxic effects of dioxin. Studies had reported a decrease in thymus weight, total lymphocyte number, and total leukocyte number. Immunologic effects from exposure to dioxin had also been reported in mice (6).
Dioxin produced anomalies in the fetus, including cleft palate and hydronephrotic kidneys in mice and internal organ hemorrhage in the rat. In studies of monkeys exposed to dioxin through diet, spontaneous abortions occurred in two-thirds of the monkeys. Rats fed high doses of dioxin had decreased fetal survival. Medium doses affected litter size and fetal and neonatal survival. The
Figure 2-2.—Availability of information on the health effects of dioxin (animal data).
lowest doses produced dilated renal pelvises, decreased fetal weight, and changes in the gestational index.
Dioxin produced mostly negative results in tests for genotoxicity; however, there were a few positive responses, which suggested that dioxin was genotoxic. Dioxin had been demonstrated to be an animal carcinogen in both rats and mice. In Swiss mice, females, but not males, developed skin tumors following treatment with dioxin. There had also been mixed evidence that dioxin was a tumor promoter.
In summary, at the time the Dioxin Subregistry was established, there was a lack of substantive information on the health effects of dioxin following nonoccupational exposures. The Dioxin Subregistry was created to aid in addressing this data gap.
SECTION 3
POPULATION SELECTION AND SITE DESCRIPTIONS
BACKGROUND
According to Kay (7), in the 1960s Hoffman-Taft in Verona, Missouri, produced a defoliant composed of 2,4-dichlorophenol and 2,4,5-trichlorophenol, which was called "agent orange" by the military due to the identifying color stripe on its container. As a result of curtailed use of the herbicide in Vietnam, the company ceased manufacturing and the facilities sat idle for a number of months. On November 18, 1969, the Northeastern Pharmaceutical and Chemical Company (NEPACCO) leased the manufacturing line to make hexachlorophene, an antiseptic. One of the by-products of this process was 2,3,7,8-tetrachlorodibenzo-p-dioxin (dioxin). Shortly thereafter, the entire plant was acquired by Syntex-Agribusiness; however, NEPACCO remained active at the plant until early 1972, when the hexachlorophene market collapsed.
During the period from February 16, 1971, through October 25, 1971, Bliss Waste Oil Service picked up six truckloads of still bottoms from NEPACCO (8). The loads, approximately 3,000 gallons each, were hauled to eastern Missouri (Figure 3-1). Almost all of this material was either sprayed directly or as a mixture with waste oil as a dust suppressant.
SITE DESCRIPTIONS
The participants in the Dioxin Subregistry represent individuals who were potentially exposed to dioxin at one of four sites (Figure 3-2) in Missouri-the Quail Run Mobile Home Park, Minker/Stout/Romaine Creek, Shenandoah Stables, and Times Beach. These sites were quite varied with respect to the type of exposure, as well as the levels of dioxin (Table 3-1). A description of each of these sites follows.
Quail Run Mobile Home Park
In January 1983, the U.S. Environmental Protection Agency (EPA) received information that the central road of the Quail Run Mobile Home Park in Gray Summit, Franklin County, had been sprayed with dioxin-contaminated oil in April 1971. Dioxin, measuring 50 parts per billion (ppb), was discovered in soil samples collected by EPA in March 1983.
Subsequent investigations revealed that the contamination had spread from the road throughout the site via tracking, wind and water erosion, and contaminated soil used as fill at nearby locations. Dioxin was detected along the entire length of the road that runs through the park, with levels ranging from 39 ppb to 1,100 ppb in composite soil samples (2,200 ppb dioxin was measured in a single noncomposite sample). It was measured at levels >1 ppb along both road shoulders, in 4 of 8 yards tested, in dust samples collected from the interiors of 21 of 31 mobile homes tested (the highest level was 11.5 ppb), in the wall insulation of at least 1 mobile home, and in the furnace air filter of at least 1 mobile home. Six soil samples in which dioxin had been detected were also tested for the presence of 132 substances considered either hazardous or priority pollutants by EPA; none
Figure 3-1.—Schematic of the fate of dioxin from the Northeastern Pharmaceutical and Chemical Company plant (8).
of the substances were detected in any of the soil samples (9). Affected areas included 28 mobile homes, a house, the roof of an underground home, soil under an equipment shed, 1,400 feet of highway shoulder, and approximately 5 acres of the park property (10).
On May 12, 1983, the Centers for Disease Control (CDC) (now the Centers for Disease Control and Prevention [CDC]) issued a health advisory which stated that the 29 families living in the park were at risk of developing adverse health effects if they remained in residence. EPA initiated an immediate removal action at the Quail Run Mobile Home Park, including the temporary relocation of the residents.
A security fence was erected by EPA during the week of May 23, 1983. Additional soil samples were collected in July to more precisely define the extent of contamination. Twenty-eight mobile homes were removed from the site and decontaminated. After restoration of the site was completed in April 1987, the park was reoccupied and renamed Fox Creek Estates (11).
Figure 3-2.—Location of the Missouri dioxin sites included in the Dioxin Subregistry of the National
Exposure Registry.
Table 3-1.—Characteristics of the Missouri dioxin sites.
| Site Name | Land Use | Maximum Dioxin Level (ppb)* | Year Contamination Level Identified | Potential Population at Risk of Exposure |
|---|---|---|---|---|
| Shenandoah Stables | Recreational | 33,000 | 1974 | 100 |
| Minker/Stout/Romaine Creek | Residential | 740 | 1983 | 100 |
| Times Beach | Residential | 980 | 1982 | 2,100 |
| Quail Run Mobile Home Park | Residential | 2,200 | 1983 | 150 |
*ppb = parts per billion; peak dioxin levels are reported in soil.
Minker/Stout/Romaine Creek
The Minker/Stout/Romaine Creek site, located in Jefferson County, consists of five separate areas, the Minker residence, the Stout residence, Romaine Creek, the Cashel residence, and the Sullins residence. These sites were contaminated when contaminated fill material, originating at the Bubbling Springs Arena, was deposited there in the early 1970s. Because of extensive erosion and runoff, the area of contamination expanded to the point where it was impractical (or impossible) to allow residents to pursue normal activities without significant risk of exposure. As a result of these facts, and the time and complexity involved in developing and implementing a remedial action plan for these sites, EPA offered to relocate the residents (4).
Bubbling Springs Arena
Bubbling Springs Arena is a recreational area on Romaine Creek Road that was used by members for horseback riding and swimming. A 0.6-acre outdoor horse arena was located on the property. The arena was named after Bubbling Springs, which surfaces upstream from the arena and forms a branch that runs parallel to the arena. This branch is a tributary of Romaine Creek.
The arena was sprayed twice in June 1971 with waste oil that contained dioxin. Approximately 2,500 gallons of oil were sprayed on the arena, and then the arena was disked to a depth of approximately 8 inches to mix the oil with the soil (12). Shortly thereafter, six horses became ill and died. At the time, the cause of death was undetermined. As a precaution, however, the owners had the arena floor excavated. The excavated soil was used as fill in nearby residential areas (7).
Samples from the arena were collected by EPA in May 1982 and February 1983. The highest level of dioxin (97 ppb) was detected in a composite sample from the perimeter of the horse arena. The dioxin had migrated and, in addition to being found in the horse arena, was also detected in soils on lands located between the horse arena and Bubbling Springs Creek, and in the pasture area.
Minker Residence
In March 1973, before the soil was analyzed for contaminants, a private contractor used approximately 20 truckloads of soil excavated from the Bubbling Springs arena to fill a steep ravine on the south end of the Minker house. EPA sampled the site in 1982.
The site contamination resulted when the fill was placed in a natural drainageway on a steep slope in a headwaters area of Romaine Creek. Natural erosion, wind transport, and vehicular and pedestrian movement resulted in documented contamination from 0 to 360 ppb as far away as 6,000 yards downgradient in Romaine Creek and at other residences near the site. On the site itself, sampling results ranged from 85 to 740 ppb (8).
Because CDC had determined that long-term exposure to dioxin at >1 ppb presented the possibility of excessive health risks to residents in the contaminated area, eight families were offered the opportunity for permanent relocation.
Stout Residence
The Stout portion of the Minker/Stout/Romaine Creek site became contaminated when dioxin-contaminated soil was used as fill material to create a level surface for two mobile home pads. It is believed that approximately 850 cubic yards of dioxin-contaminated fill material were used as part of a much larger amount of fill required to create this level surface. The site is on the side of a ridge with a natural slope of 30 degrees. The presence of dioxin was confirmed in the gully behind the property (7).
Sample results showed dioxin at this site at levels from 1 to 272 ppb. The CDC health advisory for Minker also covered this site. Three families were offered permanent relocation and moved from the site.
Sullins Residence
In 1972, the Sullins family removed a large tree from a corner of their property, located on Romaine Creek Road, and arranged for a private contractor to deliver about 14 cubic yards of fill dirt to place in the hole. The contractor advised the Sullinses that the soil was from the Bubbling Springs arena and was being removed because it had caused problems with the horses' hoofs; the Sullinses were also told that grass would probably not grow on the site because the soil had been sprayed with contaminated waste oil (7).
EPA sampled the site in December 1982, and results showed dioxin levels ranged up to 99.2 ppb. Additional samples taken in March 1983 showed contamination in the fill area as high as 820 ppb. Access to the fill area was restricted.
Cashel Residence
When Bubbling Springs arena soil was being excavated, Cashel obtained two loads of dirt which he deposited in his backyard, located on Romaine Creek Road. He reported that the dirt was "very smelly". City officials subsequently advised him not to eat vegetables grown in and around the excavated soil. Mr. Cashel reported that no vegetables grew in the soil and that two apple trees died. He scraped up the dirt and piled it at the rear of his backyard. Vegetation subsequently covered most of the contaminated area. The residence is on a ridge line between two forks of Romaine Creek. The grade to the creek is very steep and erosion could have spread the contamination (7).
EPA sampled the site in January 1983, and results showed dioxin levels from 10 to 100 ppb. Additional sampling done in April showed no higher levels of contamination. Access to the fill area was restricted.
Romaine Creek
The Minker residence is located on a ridge at the top of the watershed. One of the biggest problems at the Minker site was the severe erosion that caused considerable contamination in Romaine Creek. Dioxin levels in the creek were reported up to 270 ppb, and dioxin was detected in sediments 2 feet below the surface of the creek bed. Furthermore, bottom-feeding whole fish taken from the Meramec River near the mouth of Romaine Creek showed levels of dioxin >78 parts per trillion (ppt).
Shenandoah Stables
Shenandoah Stables are located on a 6.9-acre tract on Rural Route 1, Moscow Mills, Lincoln County. A 78.5- by 189-foot enclosed arena was sprayed with approximately 2,000 gallons of waste oil containing up to 350 parts per million (ppm) dioxin on May 25, 1971 (8). Shortly thereafter, birds and insects began dying; by August 17, 1971, 12 cats and 4 dogs had died. Of 125 horses exposed to the arena from May 1971 through January 1974, 47 horses died, 25 foals were either stillborn or died shortly after birth, and 23 mares aborted. Some foals were born with deformed legs. A daughter of the stable owner also became sick and was hospitalized for one week. She lost approximately 50% of her body weight before recovering (7).
The stables are in the floodplain of Crooked Creek, a tributary of the Cuivre River. A ditch allows water to flow from the stables to Crooked Creek. The stable area was flooded in December 1982 and in April 1983. Samples collected from the arena floor in August 1971 and analyzed in July 1974 contained from 31,800 to 33,000 ppb dioxin and from 1,350,000 to 1,590,000 ppb polychlorinated biphenyls. In October 1971, 12 inches of soil was excavated from the arena floor and replaced with clean soil. The contaminated soil was put on the southbound lane of Highway 61, which was under construction at the time.
An estimated 2 feet of soil was removed from the arena floor again in April 1972 and deposited in a slough located approximately 75 feet southeast of the stable (12). Trees in the area died the next year. Clean soil and gravel were placed over the fill and some grass, subsequently, grew there. Arena soils were collected for analyses for a second time in August 1974 and no dioxin was detected.
Composite soil samples were collected for a third time from the surface of the arena floor to a depth of 30 inches, and soil samples from around the stable were collected again on May 26, 1982, and analyzed for dioxin. Dioxin levels in these samples ranged from 1.4 to 130.0 ppb. Soils in the slough area where the April 1972 arena floor material was placed contained a maximum concentration of 1,800 ppb dioxin (12). Dust samples taken inside the arena showed contamination from 80 to 100 ppb. Based on this information, the arena was closed.
Times Beach
Times Beach, located in St. Louis County, is an urban, residential setting situated in the floodplain of the Meramec River and has a history of recurring floods. In 1971, all but 5 of the streets of the town, which covered approximately 413 acres, were sprayed for dust control with dioxin-contaminated oil.
An EPA field investigation began in Times Beach in November 1982. Soil samples were collected in late December, and dioxin contamination was confirmed at varying concentrations along the shoulders and in the drainage ditches of streets throughout Times Beach. The population of approximately 2,100 was, therefore, potentially exposed to concentrations of dioxin along roadways ranging from <1 to 980 ppb (4).
Before the laboratory analyses of the samples were completed, a flood struck a major portion of the town and led to the near total evacuation of the community. The flooding raised the possibility of movement of dioxin-contaminated soil from washed-out roadways into the debris and, conceivably, the buildings. Upon receiving the results of the initial laboratory analyses, and while awaiting further environmental sampling to clarify the location, extent, and level of contamination, CDC issued an advisory on December 23, 1982, recommending that the evacuated residents should not return to the town and that clean-up efforts be halted (or performed with full protective gear only on an emergency basis) until additional environmental sampling could delineate the extent of dioxin contamination. This advisory, unlike those at other sites, was issued on an emergency basis to temporarily prevent the repopulation of the town, especially since the main activity of the returning individuals would have been intensive cleanup of the potentially contaminated muddy soil and debris (4).
EPA conducted a second round of soil testing and analysis in January and February 1983. Post-flood environmental sampling results available in February 1983 showed only limited movement of dioxin-contaminated soils; however, based on the extensive amount of contaminated soil in and near the roadbeds, and the continued threat of flooding, and in consideration of the available remedial strategies, EPA decided to offer a phased plan of permanent relocation to the inhabitants of this affected area. On February 22, EPA announced the transfer of $33.1 million from Superfund monies to the Federal Emergency Management Agency (FEMA) to provide a complete buyout of Times Beach businesses and residences. On March 23, the governor of Missouri signed legislation appropriating the required 10% state matching funds. On June 7, FEMA and the state executed an agreement to provide permanent relocation of the affected residents (7).
POPULATION SELECTION
According to Roberts and Bagby (3), about one-third of the dioxin-contaminated sites in Missouri had levels exceeding 100 parts per billion (ppb) dioxin, and half of these were in residential areas. Because they were not uniform in geography, topography, geology, and characteristic land use, the sites varied widely in their potential for human exposure, thus presenting difficulties in public health decision making. Sites with high levels of contamination and those in areas of frequent and regular access constituted the greatest public health risk; however, at other sites, dioxin contamination was in clearly circumscribed areas, at subsurface depths exceeding 15 feet, under paved areas, or in areas with limited land use. All of these factors were considered in assessing the risk of exposure for an estimated 5,000 persons who lived in the contaminated areas from 1971 through 1983.
The Missouri Central Listing (MCL) was developed in order to enable public health agencies to keep in touch with and locate these potentially exposed individuals for educational purposes or possible epidemiologic or clinical followup, or both (4). Baseline and identifying information was collected using the Health Effects Survey (HES) questionnaire constructed to elicit, along with health information, information on exposure risk (that is, residential history; type, frequency, and duration of recreational activities; and occupational history). The questionnaire was administered to individuals or their relatives believed to be at risk of exposure to dioxin. Risk was evaluated based on proximity of residences to a site, possibility of occupational exposure to those working at a site, or frequency of activities on or around a site. However, the criteria for determining what constituted a potential exposure to dioxin were very general and essentially allowed each registrant to define his or her exposure.
For health studies, a subpopulation meeting a more rigorous definition of exposure with explicit criteria to identify dioxin exposure was chosen. This subpopulation comprised people who reported having lived or worked for at least 6 months in areas contaminated with dioxin at levels >100 ppb or for at least 2 years in areas contaminated with dioxin at levels from 20 to 100 ppb, or anyone who participated at least once a week for 6 months in activities that involved close contact with soil contaminated with dioxin at levels >100 ppb or at least once a week for 2 years in activities that involved close contact with soil contaminated with dioxin at levels from 20 to 100 ppb. Activities included gardening, field or court sports, and playing in soil (4).
The 337 persons identified as eligible for the Dioxin Subregistry are a subset of the 2,101 people who are listed on the larger MCL and were selected for the four dioxin-related health studies that were conducted from 1983 through 1986 by the Missouri Department of Health (MDH) in conjunction with the CDC. These studies were the (1) pilot epidemiology study, (2) Quail Run medical epidemiology study, (3) Quail Run follow-up study, and (4) adipose (fat) tissue study. The participants of these past dioxin health or epidemiologic studies served as the specific population from which members of the Dioxin Subregistry were selected. The Agency for Toxic Substances and Disease Registry (ATSDR) used this criterion for subregistry eligibility because people in each of the four studies were residents of housing units where there was documentation strongly suggesting exposure to dioxin. All people who participated in one or more of the four studies were asked to participate in the Dioxin Subregistry. Since the studies form the basis from which Dioxin Subregistry participants were selected, each will be briefly reviewed. The results of the studies are included in Section 6.
Pilot Epidemiology Study
In 1983, CDC selected 130 persons for inclusion in a pilot epidemiology study from 800 individuals or their relatives believed to be at risk of exposure to dioxin because they lived near, worked at, or frequently participated in activities near a contaminated site based on MCL data. A high-risk group was selected from this pool based on considerations of duration of exposure to, types of activities at, and extent and magnitude of contamination at dioxin-contaminated areas. In general, the 82 persons in this group had the longest periods of exposure (up to 12 years) to the highest levels of soil contamination (as high as 33,000 ppb) or reported having the most frequent high-soil-contact activities in these highly contaminated areas. The purpose of this selection was to maximize the likelihood of identifying individuals with potential dioxin-related health effects.
A group of "low-risk" individuals was selected mostly from the 800 respondents with the lowest risk of exposure based on the type of exposure site, age, sex, race, and socioeconomic status characteristics in a 2:1 ratio of HIGH:LOW risk subjects. This group comprised 40 persons who had not reported any access to or regular high-soil-contact activities in any known contaminated areas.
The two groups were compared using their responses on the HES questionnaire and results of a clinical protocol composed of physical, neurologic, dermatologic, hematologic, immunologic, and liver function testing (13).
Quail Run Medical Epidemiology Study
For current and former residents of the Quail Run Mobile Home Park, a comprehensive investigation was conducted that used the HES questionnaire, as well as medical and laboratory tests (9). This population was selected for study because of high levels of dioxin contamination found throughout the environs of the mobile home park, including inside many of the homes. An exposed group of 154 persons and a comparison group of 155 persons participated in the study. These people were evaluated under a protocol similar to that used in the pilot epidemiology study. In addition, more specific tests of neurobehavioral parameters (the World Health Organization's core battery for field studies of people potentially exposed to neurotoxins); quantitative tests of tactile, vibratory, and thermal sensations; and additional laboratory tests (serum IgG and creatine assays, urine cultures, assays of cytotoxic T-lymphocyte production, and liver function tests of microsomal enzyme induction) were used.
Quail Run Follow-up Study
In 1986, a follow-up study was carried out on persons from the Quail Run epidemiologic study who were anergic or relatively anergic (14). All exposed and unexposed participants from the first study who did not respond to any of the delayed-type hypersensitivity skin tests antigens ("anergic") or who responded to only one antigen ("relatively anergic") were contacted and asked to return for repeat testing. Of those contacted, 28 exposed and 15 unexposed persons were enrolled in the follow-up study.
A medical, residential, and occupational history was taken on each participant to update medical and exposure information collected during the first study. Complete blood cell counts with differentials and quantitative serum immunoglobulin tests were performed. In vitro lymphocyte tests, including T-cell surface markers (T3, T4, and T8), lymphocyte proliferative responses to mitogens and tetanus toxoid, and allogenic T-cell cytotoxicity, were conducted. Finally, the Multisite CMI skin test was conducted.
Adipose (Fat) Tissue Study
HES questionnaires were used to identify 400 persons as having had high-risk exposures by the criteria mentioned previously for health studies. A total of 59 persons agreed to participate in the adipose tissue study, in which each participant donated approximately 20 grams of subcutaneous adipose tissue from the anterior abdominal wall. These samples were evaluated for the presence of dioxin and compared with similar samples from a control group.
Summary
In summary, the participants in these studies were not mutually exclusive. People selected for the adipose tissue study included 11 persons who were also part of both Quail Run studies, 5 persons from the pilot epidemiology study, people involved in occupational exposure (stable workers from several horse arenas), and people involved in recreational exposure. The total number of participants in the four studies was 353, representing 337 unique individuals.
Of the 337 persons identified as eligible, contacted, and asked to participate in the Dioxin Subregistry, 246 agreed to participate; 4 deceased eligible persons were identified and also included. The remaining people either chose not to participate (17 persons), were in litigation over the site and chose not to participate (6 persons), were incapacitated and no proxy available (3 persons), were not at home after 8 attempts to contact (11 persons), or ineligible (1 person). ATSDR was unable to locate 49 eligible persons.
ON-SITE ACTIVITIES
In the summer of 1989, data collection for the Dioxin Subregistry began with telephone interviews (face-to-face interviews had been conducted for previous health studies) with potential subregistry members from Missouri. People who had participated in one of the four previously mentioned Missouri health studies were considered to be "potential" registrants.
Preparations for on-site data collection included a mailing of an information packet to potential registrants. The mailing consisted of a question-and-answer brochure that contained information about ATSDR, the National Exposure Registry, the Dioxin Subregistry, and the chemical dioxin. It also included a chapter from the ATSDR Toxicological Profile for Dioxin, which summarized all that was known about the association of adverse health outcomes and exposure to dioxin (6). This information packet was sent to residents before the data collection effort began. Soon after the mailout, interviewers began telephoning potential registrants.
A computer-assisted telephone interview (CATI) system was used to collect data. Each eligible person or a proxy for that person was administered the National Exposure Registry questionnaire that included a set of questions about health conditions that the registrant currently had or had ever had and that had been either confirmed or treated by a health practitioner. Each time the respondent reported the presence of one of these health conditions, a set of follow-up questions was asked about the date of first treatment by a physician, current treatment, prescribed medication, and hospitalization related to the condition. Demographic and occupational information was also collected.
Information on deceased eligible persons was obtained from a knowledgeable proxy (usually the spouse) in order to request a death certificate from the appropriate state office. Information on cause of death, along with other pertinent information, was extracted from the death certificates and coded as copies of death certificates were obtained from the states.
The first followup was conducted in 1990; the second followup was conducted in 1991. In both cases, a letter was mailed to registrants announcing the upcoming activity. Interviews were conducted using the same questionnaire and administered using a CATI system.
For the baseline activities, information was obtained for 250 persons (246 living and 4 deceased); for the first followup, 226 persons (221 living and 5 deceased); and for the second followup, 208 persons (203 living and 5 deceased). The response rate (the percent of those eligible who were contacted and asked to participate) for the baseline was 91%, for the first followup, 96%, and for the second followup, 96%.
SECTION 4
COMPARISON OF DIOXIN SUBREGISTRY DATA
AND NATIONAL DATA
REGISTRANT DESCRIPTIVE DATA
At baseline there were 250 members of the Dioxin Subregistry-246 living and 4 deceased. Of the 246 living registrants, 207 (84.1%) completed the interview themselves; 39 (15.9%) required a proxy because of age or a disability. Death certificates were obtained for the four deceased members; however, because of the small number, the deceased component of the population was not included in the statistical analyses and is not discussed further in this report. Tables 4-1 and 4-2 contain information on the characteristics of the 246 registrants alive at the time of the baseline interview. Information from all sites is included in these tables.
Table 4-1 shows that 47.2% of the registrants were female. The table also shows that 41.9% of the registrants were from 25 through 44 years of age and 13.4% were under the age of 18 years. One registrant refused to answer the age question and, therefore, was excluded from all analyses involving age. The Subregistry data were collected in 1989 and were limited to those who participated in the health studies conducted during or prior to 1985; therefore, any registrants less than 5 years of age were excluded. About 98% of the Dioxin Subregistry population were white, indicating an overall overrepresentation of whites compared with the general U.S. population.
Of the registrants 25 years of age or older, 73.4% had at minimum a high school diploma. More females than males had completed high school (52.9% versus 43.3%) or some college (19.5% versus 15.5%). However, a higher percentage of males (10.3%) than females (5.8%) had completed college or some post-college education. (See Table 4-1.)
Of the registrants 19 years of age or older, 73.4% were employed either full- or part-time at the time of the interview; this percentage varied substantially between males and females: 81.1% for males versus 64.6% for females.
Table 4-2 provides data on the registrants' use of tobacco products. Of the 212 registrants 18 years of age or older, 38.2% reported being current cigarette smokers (smoked more than 100 cigarettes, smoked at time of interview); 43.4% of females and 33.6% of males were current smokers. More females than males (37.4% versus 29.2%) reported having never smoked cigarettes. A small number of the registrants reported ever using other tobacco products: pipes (10.8%), cigars (10.4%), snuff (3.8%), and chewing tobacco (9.5%). The smoking rates of the registrants were compared with national rates, and the results are reported later in this section.
RATIONALE FOR THE COMPARISON OF THE
DIOXIN SUBREGISTRY DATA WITH NATIONAL DATA
This section includes comparisons of Dioxin Subregistry data with data from national surveys. These comparisons are consistent with the Registry objectives and goals as stated in the Policies and Procedures Manual (1); one of the goals is to provide a preliminary assessment of the extent of excess
Table 4-1.—Descriptive data for living registrants (all sites).
| Variable | Males | Females | Total | |||
|---|---|---|---|---|---|---|
| N | (%) | N | (%) | N | (%) | |
| Sex | 130 | (52.8) | 116 | (47.2) | 246 | (100.0) |
| Age (years) £17 18-24 25-44 45-64 ³65 Missing | 17 16 51 37 9 0 | (13.1) (12.3) (39.2) (28.5) (6.9) (0.0) | 16 12 52 25 10 1 | (13.8) (10.3) (44.8) (21.6) (8.6) (0.9) | 33 28 103 62 19 1 | (13.4) (11.4) (41.9) (25.2) (7.7) (0.4) |
| Education (³25 years of age) Not high school graduate High school graduate Some college College graduate or more | 30 42 15 10 | (30.9) (43.3) (15.5) (10.3) | 19 46 17 5 | (21.8) (52.9) (19.5) (5.8) | 49 88 32 15 | (26.6) (47.8) (17.4) (8.2) |
| Occupational Status (³19 years of age) Currently employed Previously employed Never employed | 90 21 0 | (81.1) (18.9) (0.0) | 62 33 1 | (64.6) (34.4) (1.0) | 152 54 1 | (73.4) (26.1) (0.5) |
adverse health conditions, if reported by Dioxin Subregistry members. Another objective is to generate-rather than test-hypotheses about dioxin exposure and health outcomes.
In addition to a comparison of Subregistry health data with national health data norms, this section includes a comparison of registrant demographic and smoking data with national data. These comparisons of demographic characteristics and smoking rates indicate the extent to which Dioxin Subregistry members are similar to the general population. Such comparisons are important because certain demographic characteristics and smoking are known to be correlated with or are probable causes of adverse health conditions. What must be kept in mind when carrying out the statistical analyses and interpreting the results is the small number of registrants. The statistical power (the power to reject no difference when the converse is true) is substantially lowered when the sample size is this small.
Table 4-2.—Tobacco use data for Dioxin Subregistry registrants 18 years of age or older (all sites).
| Variable | Males | Females | Total | |||
|---|---|---|---|---|---|---|
| N | (%) | N | (%) | N | (%) | |
| Sex | 113 | (53.3) | 99 | (46.7) | 212 | (100.0) |
| Cigarettes* Current smoker Ex-smoker Never smoked | 38 42 33 | (33.6) (37.2) (29.2) | 43 19 37 | (43.4) (19.2) (37.4) | 81 61 70 | (38.2) (28.8) (33.0) |
| Pipes Current smoker Ex-smoker Never smoked | 2 21 90 | (1.8) (18.6) (79.6) | 0 0 99 | (0.0) (0.0) (100.0) | 2 21 189 | (0.9) (9.9) (89.2) |
| Cigars Current smoker Ex-smoker Never smoked | 4 18 91 | (3.5) (15.9) (80.5) | 0 0 99 | (0.0) (0.0) (100.0) | 4 18 190 | (1.9) (8.5) (89.6) |
| Snuff Current user Ex-user Never used | 3 5 105 | (2.7) (4.4) (92.9) | 0 0 99 | (0.0) (0.0) (100.0) | 3 5 204 | (1.4) (2.4) (96.2) |
| Chewing Tobacco Current user Ex-user Never used | 8 11 94 | (7.1) (9.7) (83.2) | 0 1 98 | (0.0) (1.0) (99.0) | 8 12 192 | (3.8) (5.7) (90.6) |
*Current smoker - 100 or more cigarettes in lifetime; smoked at time of interview.
Ex-smoker - 100 or more cigarettes in lifetime; did not smoke at time of interview.
Never smoked - Smoked less than 100 cigarettes in lifetime.
SELECTION OF THE NATIONAL COMPARISON POPULATION
Dioxin Subregistry baseline data were compared with data obtained from the 1989 National Health Interview Survey (NHIS)(15). The NHIS data were used in the comparison of both demographic and health data. The 1989 NHIS survey did not include questions about cigarette smoking for the subpopulation that was used for the data analyses, the results of which are presented in this report. Hence, smoking could not be included as a factor in the models used for the NHIS statistical comparisons. Smoking rates for the Dioxin Subregistry were compared with national rates derived from several sources other than the NHIS database.
An appropriate subset of the NHIS database was selected for comparison with the National Exposure Registry (NER). The NHIS population is an appropriate comparison population because it is a representative subset of the residential, noninstitutionalized U.S. population, the population of interest for comparisons of the health status of the National Exposure Registry members. As of 1985, a stratified, multistage cluster sample design was used in the NHIS to obtain a representative sample of the target population; this information was used to create representative national norms. The weighting factors provided by the National Center for Health Statistics (NCHS) (15) were applied when using the data; otherwise, because of the NHIS sample design, the estimates would have been biased. The subset of the 1989 NHIS file used for the selected comparisons in this report comprised 116,929 respondents. The NHIS file, like the Registry file, is composed of self-reported data; the data collection instruments used were very similar. Because of the similarity of the data collection instrument used by the NHIS and the National Exposure Registry, the NHIS data were appropriate for the calculation of selected prevalence and period prevalence figures and could be used for exploratory comparison with Registry data for health outcomes. However, the mode of data collection differed: the NHIS data were collected using face-to-face interviews; the Dioxin Subregistry data by telephone interviews.
The Dioxin Subregistry sites and members are located primarily in the midwestern United States (in the Times Beach, Missouri, area), with the remainder located throughout the United States. Despite the regionality of the Dioxin Subregistry when compared with the NHIS, ATSDR's review of the regional rates for selected outcomes found no definitive evidence indicating that the overall health status of midwesterners differed significantly from that of the general U.S. population; in particular, there did not appear to be a regional difference for the outcomes reported in excess by the Dioxin Subregistry population. Rates of occurrence for selected chronic conditions are listed by geographic regions-Northeast, Midwest, South, and West-in an NCHS publication (16). Of the five health conditions reported statistically significantly more often by the Dioxin Subregistry members, none were higher for the Midwest region. It does not appear, therefore, that regional differences accounted for the increased registrant reporting found for some health conditions.
METHODS OF DESCRIPTIVE VARIABLE COMPARISONS
Demographic Characteristics
The NHIS and Dioxin Subregistry samples were compared in terms of four demographic characteristics: sex, age, race, and education level. Each of these variables and cigarette smoking are potential correlates, either directly or indirectly, with health status. Therefore, it was necessary to explore the comparability of the NHIS and Dioxin Subregistry data files for these variables. The procedures used to carry out the file comparisons follow.
Sex
The sex distributions were compared for the total populations on an age-specific subpopulation basis. Each age-specific proportion (of males or females) in the Dioxin Subregistry was compared with the corresponding standard proportion derived from the NHIS using a one-sample test for a binomial proportion (17); the NHIS proportion was used as the null value. When the normal approximation did not hold, two-tailed mid-p values were computed (17-19).
Age
Age was divided into seven categories. As discussed previously in this section, the fact that the exposure ended or occurred before 1985 created a minimum age for registrants. Comparability of sex-specific age distributions between the NHIS and Dioxin Subregistry members was assessed by performing chi-squared goodness-of-fit tests. Expected counts were generated by applying the NHIS proportions for each sex-age category to the appropriate Subregistry totals.
Race
Race is an established correlate of socioeconomic status (20) and health status (21). National data indicate that nonwhites have lower rates for cigarette smoking (22). For these reasons, race is a potential independent variable for the comparisons of health status and smoking rates. As was discussed previously in this section, there were only five nonwhites (2%) in the Dioxin Subregistry; because of this small number and because there are no known health effects associated with the race reported by the nonwhite registrants, all registrants were included in the analyses. The NHIS comparison group was restricted to those reporting race as white.
Education Level
The descriptive analyses included comparisons for education level (the highest level attained as reported by the respondent); a three-category ordinal variable was created. The categories used were 0 through 11 years of education, 12 years of education (or the equivalent of a high school diploma), and 13 or more years of education. For the these analyses, education was used as a surrogate for socioeconomic status; because this is difficult to interpret for school-aged children, analyses incorporating the variable education level were restricted to those 19 years of age or older. The sex-specific education distributions for registrants and NHIS members were compared using chi-squared goodness-of-fit tests.
Cigarette Smoking
Rates for current and past smoking behavior were compared with national data from several independent sources. For the analyses, a current smoker was defined as someone who reported being a smoker at the time of the interview and who had smoked at least 100 cigarettes in his or her lifetime. An ex-smoker was defined as someone who had ever smoked at least 100 cigarettes during his or her lifetime. People who were classified in the "ever smoked" category included both current and ex-smokers.
As discussed earlier, the 1989 NHIS file, the data file used for comparing health rates, did not contain complete smoking information for the subpopulation used in the data analyses. Therefore, the variable "smoking" could not be included as an independent factor in the statistical models. However, although smoking data were not collected for all respondents, information was collected on a sufficiently large segment (although not all) of the 1990 NHIS (23) population to calculate meaningful national rate estimates that could be used to compare with the Subregistry rates. In addition to these rates, the Subregistry rates were also compared with the national rates derived from the Behavioral Risk Factor Surveillance System (BRFSS) (24) and the 1990 National Household Survey on Drug Abuse (NHSDA) (22).
METHODS OF COMPARING HEALTH OUTCOMES
Question Comparability
The reporting rates for the Dioxin Subregistry and NHIS data were compared for each health condition. One of the factors determining the validity of the results of these comparisons was the comparability of NHIS and Dioxin Subregistry health condition questions; therefore, the comparisons were preceded by an assessment of the question comparability. The health condition questions differed in three respects: restrictions on the source of diagnosis; the time frame of occurrence or treatment; and, in some cases, the wording of the health condition. A discussion of the potential variability for the health condition questions follows. The NHIS health-related questions are presented in Appendix A and the Dioxin Subregistry health-related questions in Appendix B.
Source of Diagnosis
Dioxin Subregistry questions about health conditions specified that the source of diagnosis must be a physician or other medical provider. This qualification was intended to minimize self-diagnoses or the biased reporting of health problems by registrants that might have occurred because of an increased awareness of health due to the registrants' knowledge of the exposure and the publicity related to that exposure. The NHIS questions did not include any type of limitation or qualification concerning the source of diagnosis. Therefore, if all other factors were similar or equal, an increased reporting might be expected by NHIS respondents when compared with the Dioxin Subregistry registrants.
Time Frame
The Health Information section of the Baseline questionnaire for the Dioxin Subregistry inquired about diagnosis or treatment of conditions in one time frame-ever had (subject's lifetime). The question was phrased "Has a physician or other medical provider ever told you/SUBJECT that you/he/she had or treated you/SUBJECT for ... [condition]?"). Respondents who reported yes to this question were then asked whether they or the subject was ever treated for the condition, date of first treatment for the condition, and whether they were currently being treated for the condition.
The NHIS questionnaire focused on three different time frames. Depending on the specific condition, respondents were asked if they ever had, had within the past year, or currently had (at the time of interview) the condition. Dioxin Subregistry baseline questions about health conditions asked about diagnoses of or treatment for conditions from the point of birth through the date of the interview ("Has a physician or other medical provider ever told you/SUBJECT that you/he/she/ had or treated you/SUBJECT for CONDITION?"). Only one time frame was addressed: ever had (subject's lifetime). Respondents who reported "yes" to this question were also asked whether the subject was ever treated for the condition, when the subject was first treated for the condition, and whether the subject was currently being treated for the condition.
The NHIS questionnaire included questions that focused on three time frames-ever had the condition, had the condition within the last 12 months, or currently had the condition. With the exception of heart disease, only one time frame was used to create a reponse rate for any given health condition. The NHIS data file heart disease rate reflected a composite of responses to heart-related questions that were asked in both the "ever" and "12-month" time frames. The NHIS questionnaire asked whether respondents had ever had the heart conditions rheumatic, congenital, or coronary heart disease; angina pectoris; myocardial infarction; or any other heart attack; in the last 12 months time frame for damaged heart valve, tachycardia or rapid heart, heart murmur, or other heart trouble. In addition, for some heart-related questions, a positive response was discarded if the respondent did not answer positively to one or more other selected questionnaire items (16). A comparable heart response rate could not be created for the NER file and, therefore, a comparison with national norms could not be made for the heart condition variable.
For the other health conditions, the time frames were standardized to make the NHIS and Dioxin Subregistry rates directly comparable. Table 4-3 provides a comparison of NHIS and Dioxin Subregistry questions in terms of the time frame for each health condition. One NHIS health condition question, the effects of a stroke, was asked and rate calculated in the context "have you ever had". The questions and time frames for the subregistry and NHIS matched on this condition.
Eleven of the NHIS questions were asked in the time frame "in the past 12 months."For hypertension, the NHIS 12-month response rate was calculated using the "ever had" positive responses; however, the positive response was retained in the file only if the respondent also answered positively to one or more of nine other selected questionnaire items (16). This additional restriction might have reduced the NHIS response rate for this condition. For these 12 health conditions (see Table 4-3), the Dioxin Subregistry time frames for comparison were adjusted. In the Dioxin Subregistry, a health condition was defined as occurring "in the past 12 months" if (1) the reported date of first treatment was within the 12 months preceding the interview or (2) the subject was receiving treatment at the time of the baseline interview. This adjustment could have resulted, however, in an underestimation of these 12 conditions for Dioxin Subregistry data for the following reason. A year or more before the baseline interview, a registrant might have been told that he or she had (or was treated for) one of these 12 health conditions, but was not being treated at the time of the interview. If so, such a registrant would not have been included in the rates for these 12 health conditions.
Table 4-3.—Comparison of time frames for health condition questions.
| Dioxin Subregistry Conversion from "ever had" to | NHIS Version: "ever had" | NHIS Version: "in the past 12 months" | NHIS Version: "now have" |
|---|---|---|---|
| "Ever had" | Stroke | ||
| "in the past 12 months" ("ever had" and "currently have" and/or date of 1st treatment within past 12 months) |
Cancers (all), rash, anemia, kidney disease, urinary tract disorders, ulcer, liver problems, asthma, respiratory problems and allergies, diabetes, arthritis, hypertension |
||
| "now have" ("ever had" and "currently have") |
Speech impairment, hearing impairment, mental retardation |
There were three health conditions in the NHIS questionnaire that were queried in the time frame "do you now have". These conditions were speech impairment, hearing impairment, and mental retardation. The time frame for the comparable Dioxin Subregistry health conditions was adjusted by counting only registrants who reported that they were "currently receiving treatment" for one of these three conditions. Again, if all other factors were equal or the same, an increased reporting by the NHIS respondents when compared with the Dioxin registrants would have been expected.
Health Condition Specification
Dioxin Subregistry and NHIS questions were also compared in terms of the phrasing of the health conditions. As Table 4-4 indicates, some health conditions matched exactly, while others did not match as closely. An ATSDR panel of scientists and physicians determined reasonable matches for the Dioxin Subregistry health conditions and specific NHIS conditions (ICD-9 codes (25), or NHIS condition recodes (15)).
The nine health conditions in Class A of Table 4-4 either matched exactly or the Dioxin Subregistry version was inclusive of the NHIS version. That is, the NHIS wording of the health condition and the NHIS classification of the condition in the recodes were the same as or paralleled very closely the corresponding Dioxin Subregistry item. Class B included eight health conditions that did not match as closely, but were considered to be sufficiently similar for the purposes of the NHIS and Dioxin Subregistry comparisons. In only one case (urinary tract disorders) was there a potential for excess reporting by registrants when compared with the NHIS population because of the wording of the questions. For all other questions, related factors being consistent for the two populations, the question phrasing would have potentially resulted in higher reporting rates by the NHIS population.
Table 4-4.—Comparison of Dioxin Subregistry and National Health Interview Survey health questions.
| Q#* | Wording in Dioxin Survey | National Health Interview Survey Definition | NHIS Chronic Recodes† | ICD-9§ |
|---|---|---|---|---|
| Class A¶ | ||||
| 6 | Hypertension | Same | C508 | 401-5 |
| 8 | Kidney disease | Kidney stones Kidney infections Other kidney trouble |
C409-11 | 592 590 581-3 593 |
| 10 | Effects of stroke | Cerebrovascular disease | C509 | 430-8 |
| 14 | Liver problems | Liver disease, including cirrhosis | C302 | 571-2 573.0,.3-.9 |
| 15 | Asthma, emphysema, or chronic bronchitis |
Same | C601-2 C609 |
490-1 492 493 |
| 16 | Other respiratory allergies or problems such as hay fever |
Hay fever Allergic rhinitis without asthma | C603 | 477 |
| 17 | Diabetes | Same | C403 | 250 |
| 22 | Hearing impairment | Deaf - both ears Other hearing impairment |
C203-4 | X05 X06-9 |
| 25 | Mental retardation | Same | C208 | X19 |
| Class B** | ||||
| 3 | Cancer | Some cancers queried directly; others ascertained indirectly |
140-208 | |
| 5 | Skin rashes, eczema, or other skin allergies |
Psoriasis Dermatitis Dry (itching) skin |
C112-4 | 696 690-4 698.9 |
| 7 | Anemia or other blood disorders |
Anemia of any kind | C404 | 280-5 |
| 9 | Urinary tract disorders, including prostate trouble |
Disorders of the bladder (other than bladder infections) Diseases of prostate |
C413-4 | 594.1 596 600-2 (except 601.4) |
Table 4-4.—Continued.
| Q#* | Wording in Dioxin Survey | National Health Interview Survey Definition | NHIS Chronic Recodes† | ICD-9§ |
|---|---|---|---|---|
| Class B** | ||||
| 13 | Ulcers, gall bladder trouble, or stom- ach or intestinal problems |
Gallbladder stones Gastric, duodenal, or peptic ulcer Abdominal hernia Gastritis and duodenitis Disease of esophagus Other functional disorders of stomach or digestive system (not indigestion) Enteritis and colitis Spastic colon Diverticula of intestines Other stomach and intestinal disorders (not constipation) |
C301 C303-8 C310-3 C315 |
574 530-7 550-3 555 556 558 560-2 564.1 569 787 |
| 18 | Arthritis, rheumatism, or other joint disorders |
Arthritis Rheumatism Gout Sciatica (and lumbago) Intervertebral disc disorders Bone spur and tendinitis Disorders of bone or cartilage Bursitis |
C101-7 C109 |
711.0, .9 712.8-.9 714-6 720.0 721 729.0 724,.2-.3 722, 726 727.0, .2-.9 730.0-.3, .9 731.0, .2 732-3 |
| 19 | Rheumatic fever, heart disease, or other heart problems |
Rheumatic fever Ischemic heart disease Heart rhythm disorders Congenital heart disease Other select heart diseases |
C501-7 | 390 392-9 410-4 427.0-.6,.8-.9 785.0-.2 745-6 415-7 420.9 421.0, .9 422.9 423-4 425.0-.5, .9 426, 428 429.0-.6,.8-.9 |
| 20 | Speech impairment | Stammering and stuttering Other speech impairment |
C205-6 | X10 X11 |
*Question in Dioxin Subregistry questionnaire.
†Chronic Recodes, National Health Interview Survey Public Use Data Tape Documentation (15).
§ICD-9 is the International Classification of Diseases, 9th Revision, World Health Organization (WHO) (25).
¶Class A indicates questions match exactly or closely.
**Class B indicates questions are similar.
For six of the health conditions on the Dioxin Subregistry questionnaire, there were no parallel items in the NHIS questionnaire. These conditions pertained to symptoms-frequent periods of fatigue or tiredness; frequent periods of nausea; seizures, tremors, spells, or epilepsy; weakness, paralysis, or numbness in the arms or legs; frequent periods of anxiety, nervousness, or depression; and frequent or severe headaches. Unlike some of the national health surveys (26), environmental studies commonly ask about symptoms as well as health outcomes. Data on these symptoms, while not directly comparable with NHIS data (and not discussed in this report), are important in assessing impact of environmental contaminants on health and will be useful for making comparisons with past and current epidemiologic environmental studies, as well as future longitudinal studies.
The Dioxin Subregistry questionnaire was used to obtain information, via an open-ended questionnaire, on all types of cancers. The NHIS questionnaire, however, directly queried the respondents on only specific types of cancer-including skin, stomach, intestinal, colon, rectal, lung, breast, and prostate cancer. Information on other cancers was obtained indirectly by querying the respondent on hospital stays, doctor visits, and restricted activity. The NHIS question is worded, "In the last 12 months, did anyone in the family have (specific site) cancer?" The NHIS time frame restriction and the possible restriction on types of cancers reported for the NHIS file makes the results of the comparison with the Dioxin Subregistry data uncertain and the interpretation tenuous.
Statistical Analysis of Health Data
Standardized morbidity ratios were computed for each health condition. The risk ratios were calculated as the sum of the stratum-specific observed events divided by the sum of the stratum-specific expected events derived from a standard population (27). The expected number of events for a particular sex-age stratum was computed by applying the sex-age-specific prevalence or period prevalence rate from the NHIS to the corresponding sex-age-specific denominator of the Dioxin Subregistry. The 99% confidence intervals (CIs) for the ratios were generated using exact methods, with the observed counts assumed to follow the Poisson distribution (27). The sparseness of the data precluded a detailed statistical analysis of the sex-, age-, and sex-age- specific observed/expected (O/E) ratios. The O/E ratios are presented in Appendices C-1 through C-16. For conditions in which the observed and expected counts were considered to be sufficiently large, exact 99% CIs similar to those mentioned previously were computed for the sex-specific O/E ratios. No adjustments were made for multiple comparisons.
The complex sampling design of the NHIS was accounted for when computing rates (15). Numerators for a given sex-age stratum were computed by summing the cross product of the "condition weight" and the "basic final weight" for people who were asked the appropriate "condition list" and responded positively for the condition of interest. Since the presence of cancer could be ascertained from several sources within the NHIS questionnaire, stratum-specific numerators were computed by summing the "basic final weight" for all people with cancer. Denominators were obtained from the NHIS "person" records by summing the "basic final weight" for all people in a particular stratum.
RESULTS OF DESCRIPTIVE COMPARISONS
This section provides a discussion of the results of the analyses completed to assess the comparability of the descriptive data-the demographic information and smoking history- between the Dioxin Subregistry and NHIS files.
Demographics
Race
In the Dioxin Subregistry sample, 98% (n = 241) of the registrants selected their race as white. As discussed earlier in this section, because of the small number of nonwhites in the Subregistry population compared with the NHIS population, only whites are included in the NHIS comparison population used in this report.
Sex
Table 4-5 provides comparisons of the proportion of males within age categories for the Dioxin Subregistry and the NHIS populations. The proportion of females can be calculated as unity minus the proportion of males; the p values are identical when comparing the proportions for females. No statistically significant differences were found in any of the age categories examined. The 55 through 64 years of age category exhibited borderline significance; there were more males in the Subregistry than expected.
Age
The age distributions of the Dioxin Subregistry and the NHIS populations are presented in Table 4-6. These distributions were compared statistically, stratified by sex, using chi-squared goodness-of-fit tests. For males, the age distribution of the Subregistry population was not statistically significantly different from that of the NHIS population (p = 0.11). A marginally significant difference was detected for the female subpopulation (p = 0.08), with the primary disparity occurring in the 65 years of age and older category; fewer registrants fell in this category than were expected.
Education
The distributions of the variable highest level of education attained for the Dioxin Subregistry and the NHIS participants 19 years of age or older are presented in Table 4-7. These distributions were statistically compared using chi-squared goodness-of-fit tests on a sex-specific basis.
The distribution of education level for males in the Subregistry was statistically significantly different from that of the males in the NHIS (p = 0.002). The Subregistry population contained a higher percentage of male members with less than 12 years of education (28.8% versus 20.3%) than expected based on NHIS figures and a lower percentage of college educated males (27.0% versus 43.5%). For females, the Subregistry education distribution was marginally statistically significantly different from that of the NHIS (p = 0.07). There was a higher percentage of females with only a
Table 4-5.—Percentage of males within age groups.
| Age (years) | Dioxin Subregistry* | National Health Interview Survey* | p value† |
|---|---|---|---|
| 8-17 | 0.52 | 0.51 | 0.96 |
| 18-24 | 0.57 | 0.49 | 0.39 |
| 25-34 | 0.45 | 0.50 | 0.50 |
| 35-44 | 0.54 | 0.50 | 0.55 |
| 45-54 | 0.55 | 0.49 | 0.45 |
| 55-64 | 0.68 | 0.48 | 0.05 |
| ³65 | 0.47 | 0.42 | 0.62 |
*Cell values are proportions of males for the given age group.
†Two-tailed p value, one-sample test for a binomial proportion
(two-tailed mid-p values were computed when the normal
approximation did not hold).
high school education than expected based on the NHIS rates (54.1% versus 41.3%), and a lower percentage of college educated females (28.2% versus 38.1%).
Cigarette Smoking
Table 4-8 provides smoking rates for registrants 18 years of age or older. The rates for registrants were higher when compared with the national rates for both current smoker and ever smoked categories; the rates were particularly high for women. Also included in Table 4-8 are the rates calculated using the information from the 1990 NHSDA (22). The rates listed are for the total population; in most cases, for the subpopulation reporting race as white, the rates are
about 2 % higher. As when previously compared with national norms, the smoking rates for almost all age and age-sex groups are elevated for the Subregistry population; the female rates from the Subregistry are particularly elevated. The NHSDA population rate for the North Central Region (which includes Missouri) is higher than the national NHSDA levels (32.1% and 30.3 %, respectively) but are still below the Dioxin Subregistry rate (38.2%).
Information gathered under the BRFSS (24) indicated the 1989 U.S. median rate for currently smoking was 24.2% for those 18 years of age or older, and the Missouri rate was 25.9% (compared with the Subregistry rate of 38.2%). When the BRFSS respondents were restricted to those with low education (the education level for whom smoking rates are historically the highest), the rates rose to 29.4% and 30.1% for the United States and Missouri, respectively; these rates compare with the Dioxin Subregistry rate of 38.2% for all education levels.
Table 4-6.—Comparison of the National Health Interview Survey and Dioxin Subregistry age distributions.
| Age Group (years) | National Health Interview Survey* | Dioxin Subregistry† | |
|---|---|---|---|
| % of Total | % of Total | Number | |
| Males and Females | |||
| 8-17 18-24 25-34 35-44 45-54 55-64 ³65 Total | 15.0 11.4 19.7 17.0 11.8 10.5 14.7 100.0 | 13.5 11.4 19.2 22.9 16.3 9.0 7.8 100.0 | 33 28 47 56 40 22 19 245 |
| Males Only | |||
| 8-17 18-24 25-34 35-44 45-54 55-64 ³65 Total | 15.8 11.5 20.2 17.5 12.0 10.3 12.7 100.0 | 13.1 12.3 16.2 23.1 16.9 11.5 6.9 100.0 | 17 16 21 30 22 15 9 130 |
| Females Only | |||
| 8-17 18-24 25-34 35-44 45-54 55-64 ³65 Total | 14.2 11.2 19.2 16.6 11.6 10.6 16.6 100.0 | 13.9 10.4 22.6 22.6 15.7 6.1 8.7 100.0 | 16 12 26 26 18 7 10 115 |
*Cell values are percents based on weighted frequencies.
†One registrant had missing age status.
Table 4-7.—Comparison of the National Health Interview Survey and Dioxin Subregistry education distributions for those 19 years of age and older.
| Education Level (years) | National Health Interview Survey* | Dioxin Subregistry | |
|---|---|---|---|
| % of Total | % of Total | Number | |
| Males and Females | |||
| £11 12 ³13 Total | 20.4 38.9 40.7 100.0 | 24.6 47.8 27.5 100.0 | 51 99 57 207 |
| Males Only | |||
| £11 12 ³13 Total | 20.3 36.2 43.5 100.0 | 28.8 44.1 27.0 100.0 | 32 49 30 111 |
| Females Only | |||
| £11 12 ³13 Total | 20.6 41.3 38.1 100.0 | 19.8 52.1 28.1 100.0 | 19 50 27 96 |
*Cell values are percents based on weighted frequencies.
Table 4-8.—Reported rates of cigarette smoking in the Dioxin Subregistry and comparison population.
| Age Group | N | Current Rate* | Ever Rate* | ||||
|---|---|---|---|---|---|---|---|
| Subregistry | NHSDA† | NHSDA† (North Central) | Subregistry | NHSDA† | NHSDA† (North Central) | ||
| ³18 years All Males Females | 212 113 99 | 38.2 33.6 43.4 | 30.3 | 32.1 | 67.0 70.8 62.6 | 48.6 55.2 42.7 | 56.5 |
| 18-24 years All Males Females | 28 16 12 | 39.3 31.3 50.0 | 29.6 33.7 25.7 | 34.2 | 42.9 37.5 50.0 | 36.8 40.0 33.8 | 41.0 |
| 25-34 years All Males Females | 47 21 26 | 42.6 42.9 42.3 | 37.1 39.2 35.0 | 37.6 | 68.1 76.2 61.5 | 53.5 56.3 50.8 | 53.3 |
| ³35 years All Males Females | 137 76 61 | 36.5 31.6 42.6 | 24.3 27.1 21.8 | 24.4 | 71.5 76.3 65.6 | 55.6 69.3 43.6 | 58.1 |
*Rates are number per 100.
†Source: 1990 National Household Survey on Drug Abuse, Substance Abuse and Mental
Health Services Administration (22).
Also, the Dioxin Subregistry current smoking rates exceeded those of the 1990 NHIS (males, 27.9% and females, 23.4%) (23). Unlike the reporting patterns for the 1990 NHIS file and other populations representing national norms, there was an excess in reporting by females over males for most age groups among Dioxin Subregistry members. The overall NHIS rate for males of 32.7% versus 25.8% for females (32.8% versus 28.6% for the East North Central states) compares with the Dioxin Subregistry current smoking rate of 33.6% for males and 43.4% for females.
Table 4-9 presents the ever smoked rates by education level for the Dioxin Subregistry members. As expected from the literature, the general pattern was a decrease in smoking with increased education; the rates for people with no high school diploma were considerably higher when compared with the rates of people with at least a college degree. This overall trend is replicated in other reports as well (28).
Table 4-9.—Smoking rates for the Dioxin Subregistry by educational attainment.
| Age Group | Education Level Attained | |||||||
|---|---|---|---|---|---|---|---|---|
| No High School Diploma | High School Graduate | Some College | College Graduate* | |||||
| Percent† | N§ | Percent | N | Percent | N | Percent | N | |
| ³18 years All Males Females | 75.9 76.5 75.0 | 54 34 20 | 69.3 69.4 69.2 | 101 49 52 | 57.5 73.7 42.9 | 40 19 21 | 47.1 54.6 33.3 | 17 11 6 |
| 18-24 years All Males Females | 80.0 75.0 100.0 | 5 4 1 | 46.2 28.6 66.7 | 13 7 6 | 12.5 25.0 0.0 | 8 4 4 | 50.0 0.0 100.0 | 2 1 1 |
| 25-34 years All Males Females | 70.0 100.0 40.0 | 10 5 5 | 60.9 55.6 64.3 | 23 9 14 | 81.8 80.0 83.3 | 11 5 6 | 66.7 100.0 0.0 | 3 2 1 |
| ³35 years All Males Females | 76.9 72.0 85.7 | 39 25 14 | 76.9 81.8 71.9 | 65 33 32 | 61.9 90.0 36.4 | 21 10 11 | 41.7 50.0 25.0 | 12 8 4 |
*May be additional years post-bachelor degree.
†Percent who ever smoked.
§Total number of registrants in this category.
In summary, based on a comparison of smoking rates from the Dioxin Subregistry and national sources, it is possible, because of the increased rates of smoking by Dioxin Subregistry members, that smoking is a factor to be considered, particularly for females, when comparing certain reported health outcomes from the Dioxin Subregistry and 1989 NHIS files. Smoking status must be considered when interpreting findings for outcomes for which it might play a role (29).
RESULTS OF HEALTH OUTCOME COMPARISONS
Standardized morbidity ratios, the ratios of observed events in the Dioxin Subregistry to expected events based on the NHIS and standardized for age and sex, are displayed in Table 4-10 for each health outcome. A value greater than one indicates an excess of positive responses by the Dioxin Subregistry members relative to the number expected based on NHIS rates. The risk ratios
Table 4-10.—Summary of risk ratios for total population.
| Condition | Observed (Dioxin) | Expected (National Health Interview Survey) | Risk Ratio | 99% CI* |
|---|---|---|---|---|
| Significantly elevated risk ratios† | ||||
| Cancers (all) | 8 | 2.25 | 3.56 | 1.14, 8.27 |
| Anemia and other blood disorders | 12 | 3.39 | 3.54 | 1.46, 7.12 |
| Skin rashes | 29 | 16.00 | 1.81 | 1.06, 2.87 |
| Stroke | 8 | 2.11 | 3.79 | 1.22, 8.80 |
| Urinary tract disorders | 12 | 3.42 | 3.51 | 1.45, 7.07 |
| Significantly lowered risk ratio† | ||||
| Hearing impairment | 3 | 21.97 | 0.14 | 0.02, 0.50 |
| Respiratory allergies | 12 | 25.25 | 0.48 | 0.20, 0.96 |
| Nonsignificant risk ratios† | ||||
| Arthritis | 33 | 39.99 | 0.83 | 0.50, 1.27 |
| Asthma, emphysema | 15 | 23.16 | 0.65 | 0.30, 1.22 |
| Diabetes | 9 | 6.11 | 1.47 | 0.51, 3.28 |
| Kidney disease | 8 | 3.89 | 2.06 | 0.66, 4.78 |
| Liver problems | 2 | 0.85 | 2.36 | 0.12,10.94 |
| Mental retardation | 2 | 1.18 | 1.69 | 0.09, 7.84 |
| Speech impairment | 1 | 1.64 | 0.61 | 0.00, 4.55 |
| Hypertension | 29 | 26.58 | 1.09 | 0.64, 1.73 |
| Stomach problems, ulcers | 25 | 19.66 | 1.27 | 0.71, 2.09 |
*Confidence intervals.
†Statistically significantly different from 1 at the p = 0.01 level.
for several health conditions were statistically significantly elevated (greater than 1) at the p £ 0.01 level. These were anemia and other blood disorders, all cancers, skin rashes, stroke, and urinary tract disorders. Statistically significant deficits in reporting by the Dioxin Subregistry population were observed for hearing impairment and respiratory allergies.
In addition to the overall risk ratio, the sex-, age-, and sex-age-specific (O/E) ratios are presented in Appendices C-1 through C-16. Sparse data precluded a statistical evaluation of all age- and sex-age group ratios. When the O/E counts were considered to be sufficiently large, exact 99% confidence intervals were generated for the sex- and sex-age group ratios. These results are summarized in Table 4-11. Also, the small O/E counts in many of the sex, age, and sex-age strata make observational interpretation of the data somewhat tenuous. That is, the O/E ratio, the risk ratio, might be quite large, yet its components so small that any change in the numerator (the observed count in the Subregistry) would substantially alter the ratio. A detailed examination of the results for each health outcome follows.
Anemia or Other Blood Disorders
The overall risk ratio of 3.54 (12 observed versus 3.4 expected, 99% CI = 1.46, 7.12) indicated a statistically significant increase in reporting for the Dioxin Subregistry population over that expected based on the NHIS reporting rate. The sex-, age-, and sex-age-specific estimated risk ratios for anemia are detailed in Appendix C-1. The largest excess in reporting by sex was seen for males. The O/E ratio of 7.08 for males (4 observed versus 0.6 expected) was statistically significantly greater than 1.0 (99% CI = 1.19, 22.29). The ages of the four male registrants reporting anemia covered a wide range-age groups 8 through 17 years of age to 65 years of age or older. The risk ratio of 2.83 for females (8 observed, 2.8 expected) was not statistically greater than 1.0 (99% CI = 0.91, 6.57).
Arthritis, Rheumatism, or Other Joint Disorders
The overall risk ratio for arthritis, rheumatism, or other joint disorders was less than one (O/E = 0.83, 99% CI = 0.50, 1.27), indicating fewer positive reports by the Subregistry population than expected, 33 observed versus 40.0 expected. The sex, age, and sex-age group estimates for arthritis are shown in Appendix C-2. Most of the sex-age group ratios were less than one, indicating that the reporting of arthritis was generally lower in the Dioxin Subregistry data than expected based on the NHIS prevalence rates. The one notable exception was the three positive reports for females less than 24 years of age (2 for those 18 years of age or younger), which was higher than anticipated based on NHIS rates (3 observed versus 0.64 expected).
Asthma, Emphysema, or Chronic Bronchitis
The estimated overall risk ratio of 0.65 (99% CI = 0.30, 1.22) was decreased, however, not statistically significantly decreased. The stratum-specific risk ratio estimates are given in Appendix C-3.
Cancers (All)
Comparing the Dioxin Subregistry and NHIS 12-month time frame reporting rates, the overall ratio for all cancers was statistically significantly elevated at the p = 0.01 level (risk ratio = 3.55,
Table 4-11.—Statistically significant* risk ratios for total population and age, sex subgroups.
| Condition | Sex | Age | Observed | Expected | Risk Ratio | 99% CI† (exact) |
|---|---|---|---|---|---|---|
| Hearing impairment | All Male Female | All All All | 3 3 0 | 21.97 14.87 7.09 | 0.14 0.21 0.00 | 0.02, 0.50 0.02, 0.74 0.00, 0.65 |
| Rash | All | All | 29 | 16.00 | 1.81 | 1.06, 2.87 |
| Stroke | All All Male | All 45-54 45-54 | 8 3 2 | 2.11 0.17 0.08 | 3.79 17.79 24.00 | 1.22, 8.80 2.00, 65.09 1.24, 111.28 |
| Urinary tract problems | All Female Female | All All 65+ | 12 8 3 | 3.42 1.44 0.26 | 3.51 5.56 11.58 | 1.45, 7.07 1.79, 12.91 1.30, 42.39 |
| Cancers (all) | All | All | 8 | 2.25 | 3.56 | 1.14, 8.27 |
| Anemia | All Male All | All All 55-64 | 12 4 3 | 3.39 0.56 0.25 | 3.54 7.08 11.94 | 1.46, 7.12 1.19, 22.28 1.34, 43.70 |
| Respiratory allergies | All | All | 12 | 25.25 | 0.48 | 0.20, 0.96 |
*Statistically significant at the p = 0.01 level.
†Confidence intervals.
99% CI = 1.14, 8.25). There were 8 reported cancer cases-3 thyroid, 1 bladder, and 4 skin cancers (See Table 4-12); 2.5 cases were expected based on NHIS reporting rates. Appendix C-4 shows the results for the cancers reported using the NHIS 12-month period prevalence rates as comparison values. There were a number of O/E subgroup ratios greater than 1.0; however, none were statistically significantly greater than 1.0. For example, the estimated risk ratio for females was 4.25, 5 cases observed versus 1.18 expected. All five cases occurred in women 45 years of age or older, with three occurring in those 65 years of age or older. A further analysis of the thyroid cancers, which were in excess of the number expected based on the Surveillance, Epidemiology, and End Results (SEER) program (30) (the SEER rate for white females is about 7.0 per 100,000), showed that the date of first treatment occurred before 1972 for one case and in the mid-1970s for a second case; two were females and one was male. This information indicated that, in all probability, the excess thyroid cancers were not related to exposure to dioxin.
When the time frame for reporting was ever had, the registrants reported 17 cancers (see Table 4-13) versus 8 for the last 12-month time frame. Five of the additional nine cases were females who reported positively (three in the 25 to 34 years age group) to having been told they had or
Table 4-12.—Types of cancer in the Dioxin Subregistry reported by age group and sex (time frame is last 12 months).
| Age (years) | Type of Cancer* | |
|---|---|---|
| Male | Female | |
| 8-17 | ||
| 18-24 | ||
| 25-34 | Thyroid (1) | |
| 35-44 | ||
| 45-54 | Bladder (1) | Thyroid (1) |
| 55-64 | Skin (1) | Skin (1) |
| ³65 | Skin (2) Thyroid (1) |
|
*Numbers in parentheses represent
number of cases.
treated them for cervix uteri or body of uterus cancers. Also of note is that of the 17 cancers reported, only 1 reported the date of first treatment as before 1972, 3 as mid- or late-1970s, and the remainder the 1980s.
Diabetes
None of the risk ratios, overall or sex-specific, were statistically significantly elevated for diabetes. Age and sex group details appear in Appendix C-5. Increased reporting was found for groups older than 35 years of age, with most excesses in the female groups. The largest excess (O/E = 4.36, 3 reported versus 0.70 expected) occurred in females aged 45 through 54 years.
Hearing Impairment
The overall risk ratio of 0.14 (99% CI = 0.02, 0.50) reflects 3 cases observed versus 22.0 expected. This reversal (increased reporting by NHIS versus registrants), seen for both males and females, was predictable. The restriction that Subregistry cases must have been confirmed or treated by a health care provider and NHIS cases did not could have resulted in substantially less reporting by the Dioxin Subregistry participants relative to NHIS reporting rates. The stratum-specific risk ratio estimates for hearing impairment are presented in Appendix C-6. Decreased reporting was observed in the Dioxin Subregistry data for most subgroups.
Table 4-13.—Types of cancer in the Dioxin Subregistry, reported by age group and sex (time frame is ever had).
| Age (years) | Type of Cancer* | |
|---|---|---|
| Male | Female | |
| 8-17 | ||
| 18-24 | ||
| 25-34 | Thyroid (1) | Cervix uteri (3) Head, face, neck (1) Uterus (1) |
| 35-44 | Leukemia (1) | |
| 45-54 | Bladder (1) | Thyroid (1) Colon (1) Uterus (1) |
| 55-64 | Tongue (1) | Skin (1) |
| ³65 | Skin (1) | Skin (2) Thyroid (1) |
*Numbers in parentheses represent number
of cases.
High Blood Pressure (Hypertension)
The estimated risk ratios were not statistically elevated for hypertension (O/E = 1.09, 29 observed versus 26.58 expected, 99%