HEALTH SURVEILLANCE IN A COMMUNITY AFFECTED

BY ARSENIC-CONTAMINATED WATER

Lynda Knobeloch¹, Charles Warzecha¹, Marc Weisskopf^1,2, and Henry Anderson¹

¹Wisconsin Department of Health and Family Services, 1 W Wilson, Madison, WI 53703

²Epidemic Intelligence Service, Epidemiology Program Office, Centers for Disease Control and Prevention, 1600 Clifton Rd MS-D18, Atlanta, GA 30333, email requests to C. Warzecha at warzecj@dhfs.state.wi.us

Abstract

Many private drinking water wells in Wisconsin's Fox River Valley contain naturally-occurring arsenic. The Department of Health and Family Services used a self-administered questionnaire to collect information regarding the water use habits, length of residence, and health histories of 1,838 individuals. Respondents whose lifetime arsenic intakes exceeded 500 mg/L-years were approximately five times more likely to report cancer and seven times more likely to report skin cancer than respondents with lower arsenic intakes. Between 2000 and 2002 the Department plans to reevaluate arsenic exposure and health status in this region. Three factors make this re-evaluation timely. The first is an expected proposal to lower the safe drinking water standard, which will increase the number of wells deemed non-potable. The second is the dramatic increase in the exposed population since the previous study. The final, and possibly most concerning factor, is an apparent degradation of water quality over time in this area.

Introduction

In 1987, a routine feasibility study for a proposed landfill revealed high levels of arsenic in five private drinking water wells located in central Outagamie County (Stoll et al., 1995). Additional groundwater testing identified a large arsenic source in a bedrock aquifer located at the interface of the St. Peter Sandstone and Sinnipee Dolomite. This formation stretches from southern Brown County into Outagamie and Winnebago Counties and lies beneath more than 20,000 private drinking water wells (see figure 1). Subsequent sampling of 1,943 private wells in the Fox River Valley, one of the fastest growing regions of the state, found arsenic concentrations ranging from below detection to 12,000 µg/L. Approximately one-third of these wells had arsenic levels above 5 µg/L, and 68 (3.5%) of the wells had arsenic concentrations that exceeded the federal standard of 50 µg/L. Repeated testing of several domestic wells indicates that arsenic concentrations are increasing over time. This effect may be the result of an oxidation reaction that is initiated during well construction and exacerbated by water use and draw-down of the water table.

Exposure to inorganic arsenic has been associated with a variety of health problems including circulatory disorders, gastrointestinal upsets, diabetes, peripheral neuropathies, and skin lesions (NRC, 1999). Epidemiologic studies conducted in several countries (e.g., Taiwan, Japan, England, Hungary, Mexico, Chile, and Argentina) have linked arsenic-contaminated drinking water to skin cancer in exposed populations. Increased mortality from internal cancers of liver, bladder, kidney, and lung has also been reported (NRC, 1999).

The current federal drinking water standard for arsenic was developed using an estimated adult dietary intake rate of 900 micrograms of arsenic per day and an assumption that drinking water should provide no more than 10% of an adult’s daily arsenic intake. However, after the standard was established the estimate of dietary arsenic intake was reduced from 900 to 70 micrograms per day (NRC, 1983). In 1980, EPA's Cancer Assessment Group (CAG) calculated excess cancer risk values for arsenic. CAG estimated that drinking water concentrations of 22 ng/L, 2.2 ng/L and 0.22 ng/L corresponded to lifetime cancer risks of 10^-5, 10^-6, and 10^-7, respectively (Fed Reg, 1985). This assessment, combined with the reduced dietary intake estimate, has put pressure on EPA to reduce the amount of arsenic that is allowed in public water supplies. The agency is expected to propose a new standard for arsenic during the summer of 2000.

In an effort to assess the public health impact of arsenic-contaminated drinking water, the Wisconsin Department of Health and Family Services conducted a health and drinking water study in the Fox River Valley (Haupert et al., 1996). More than 600 families provided water samples for analysis and completed health surveys. This article reexamines findings from that study and discusses a follow-up study that will begin in July 2000.

Methods

The Winnebago and Outagamie County Health Departments issued press releases encouraging private well owners to submit a water sample for arsenic analysis. These announcements were published in local newspapers. Families that requested sample kits were asked to complete a health survey and sign a release form that authorized investigators to access their water test result. Most of the surveys were returned before families received their water test results, which minimized the potential for recall bias. Arsenic analyses were performed by state-certified laboratories using standard EPA-approved methods for quantification. Survey and water test data were analyzed using EpiInfo 6.01 software, developed by the U.S. Centers for Disease Control and Prevention, and SAS 8.0. The local health departments are again committed to assisting with water sample collection and health surveillance activities.

Results

995 families submitted water samples. Arsenic levels in these samples ranged from below detection to 999 µg/L. Most of the families had drinking water arsenic levels of less than 5 µg/L, and only 4% had levels that exceeded the federal standard of 50 µg/L (see Table 1). Health surveys were returned by 637 families providing water intake and health information for 1,838 individuals. Respondents ranged from <1 to 91 years in age and included 426 children.

Cancer incidence was evaluated among respondents who had lived in their homes for at least 5 years. Lifetime arsenic intakes were estimated by multiplying the drinking water arsenic level by the number of years each individual had lived in his/her home prior to the survey or cancer diagnosis. Overall cancer rates (all sites combined), and rates for non-melanoma skin cancer were positively associated with arsenic intake estimates (see Table 2). Six of the nine most highly exposed respondents reported malignancies. Each of these respondents had lived in their homes for more than 20 years and had a well water arsenic level that exceeded 100 µg/L. Cancer types reported by these individuals included skin (3), prostate, bladder, and colon.

Daily arsenic intakes were estimated for each respondent by multiplying water consumption in liters/day by the drinking water arsenic concentration in µg/L. Respondents whose estimated arsenic intakes exceeded 49 µg/day were more likely to report high blood pressure, ulcers, tremors, and unexplained hair loss than others (see table 3). Non-cancerous skin changes, numbness, headaches and gastrointestinal illness rates were not correlated with daily arsenic intakes.

Discussion

This community health survey found an increased incidence of cancer and other health problems among residents who consumed water from arsenic-contaminated wells. These findings must be interpreted cautiously, however, since the number of exposed individuals and illnesses was quite small and because the illnesses and cancer diagnoses were self-reported. Information provided by survey respondents was not confirmed by a review of medical records. Confirmation would have required a review of each respondent’s medical records since, with the exception of internal cancers, these conditions are not reportable under current state statutes. In addition, residential history and information on other sources of drinking water, e.g. water consumed at work or school, was not considered in this study.

Between 2000 and 2002, the Department plans to re-evaluate arsenic exposure and health status in this region which has grown rapidly during the past decade. This re-evaluation should provide stronger conclusions based on a significantly larger study population. In addition, the anticipated increased number of participants may allow us to study health effects at lower exposure levels that are more relevant to a proposed change in the drinking water standard. The new study will also consider health endpoints such as diabetes and cardiovascular disease that were not evaluated in the previous survey.

References

1) Stoll R, Burkel R, and N LaPlant, 1995. Naturally-occurring arsenic in sandstone aquifer water supply wells of Northeastern Wisconsin. WI Groundwater Research and Monitoring Project Summaries. WI Dept Natural Resources PUBL-WR-423-95.

2) National Research Council, 1999. Arsenic in drinking water. National Academy Press, Washington DC.

3) National Research Council, 1983. Drinking Water and Health, Vol 5. National Academy Press, Washington DC.

4) Federal Register Vol. 50, No. 219, November 13, 1985. National Primary Drinking Water Regulations.

5) Haupert, TA, Wiersma, JH, and JM Goldring, 1996. Health effects of ingesting arsenic-contaminated groundwater. Wisc Med J; 95(2):100-104.

Figure 1. Map illustrating affected region.

Table 1. Description of respondents
Characteristic	Number (%)
Age 0-17 18-64 65+ Not provided Gender Male Female Water arsenic concentration (µg/L) < 5 5-49 50+ Years in current home Not provided <5 5-9 10+ No. of 8-oz. glasses/day Not provided 0 1-2 3-4 5+	426 (23) 1040 (56) 157 (9) 215 (12) 921 (50) 917 (50) 1163 (63) 602 (33) 73 (4) 178 (10) 620 (34) 313 (17) 727 (40) 51 (3) 70 (4) 465 (25) 749 (41) 503 (27)

Table 2: Incidence and age-adjusted relative risk of cancer and skin cancer
Arsenic intake (µg/L x years in home)	Cancer incidence (%)	RR (95% CI)	Skin cancer Incidence (%)	RR (95% CI)
0-499	46/962 (5%)	1.0 (ref)	14/962 (1%)	1.0 (ref)
500-999	2/35 (6%)	1.5 (0.3-6.5)	2/35 (6%)	5.1 (1.1-24.4)
1000+	6/30 (20%)	5.2 (1.8-14.4)	3/30 (10%)	6.9 (1.7-27.2)
0-53720	54/1027 (5%)		19/1027 (2%)

Table 3: Incidence of non-cancer health effects
Health concern	Age-adjusted Relative Risk (95% CI) >49 vs < 49 µg/day
High blood pressure	1.7 (0.7-4.3)
Ulcers	1.9 (0.4-8.1)
Tremors	2.7 (0.3-20.8)
Hair loss	4.2 (1.2-14.2)