Millions of people in Bangladesh and West Bengal, India are still being exposed to high levels of arsenic in their drinking water, despite a million-dollar screening effort to distinguish safe from unsafe wells, according to laboratory data, which were reported on ES&T’s Research ASAPs website (ES020591o) and are scheduled to appear in the December 15 issue of ES&T, by Dipankar Chakraborti, head of the School of Environmental Studies at Jadavpur University in Calcutta, India, and colleagues. The new results suggest that the field kits used to measure arsenic in the region’s groundwater are unreliable and that many wells have been incorrectly labeled.

The arsenic poisoning problem began with a decision made decades ago to switch from bacteria-laden surface water to groundwater. Although infant mortality and diseases like cholera have decreased significantly as a result of this policy, the symptoms of chronic arsenic poisoning have become an all-too-familiar occurrence because of naturally high levels of arsenic in the underlying sediment.

To assess the magnitude of arsenic contamination, the World Bank, the United Nations Children’s Emergency Fund (UNICEF), the World Health Organization (WHO), and several other international aid agencies made a joint decision in 1997 to test all hand-pumped tube wells using colorimetric field kits. To date, more than a million wells have been tested using the kits. Those with arsenic levels >50 micrograms per liter (µg/L) were painted red to indicate that the water is unsafe for drinking, and those with levels <50 µg/L were painted green to indicate that the water is safe. Millions of dollars have been spent testing the wells, and millions more are slated for future screening efforts.

A young villager in Bangladesh pumps water from a well, which was painted green to indicate that the water has <50 µg/L arsenic and is safe to drink. Credit: UNICEF-Bangladesh

In the field
Several arsenic field test kits are commercially available, but thus far, most wells in Bangladesh, where the largest population is at risk, have been tested using the Merck kit, which, like all other field kits used to date in Bangladesh, is based on the mercuric bromide stain method. Also known as the Gutzeit method, the procedure involves reducing inorganic arsenic to arsine gas, which then reacts with mercuric bromide on a test strip and turns it from white to yellow or brown, depending on the concentration. By comparing the color of the test strip to a color scale provided with the kit, the amount of arsenic in a sample can be estimated. The first-generation Merck kit had a minimum detection limit of 100 µg/L; however, beginning in 2000, an improved Merck kit, with a minimum detection limit of 10 µg/L, was introduced. Despite the improvement in sensitivity, most experts agree that the kit works well for samples containing arsenic levels <10 µg/L or >100 µg/L, but not in between. In addition, there are concerns about the toxicity of the arsine gas produced by the kits (Environ. Sci. Technol. 1999, 33, 3686–3688).

Perhaps the weakest points of the Merck kit are the color scale, which, as shown in Figure 1, only has five colored standards, and the glass reaction tubes, says Jan Willem Rosenboom, the arsenic project officer for UNICEF Bangladesh. As a result, the “de facto” preferred kit for both UNICEF and the government’s response program, the Bangladesh Arsenic Mitigation Water Supply Project (BAMWSP), is now a field kit produced by the U.S.-based instrument company, Hach. According to Rosenboom, the Hach field kit is simpler to use and has an improved color scale, with more color standards in the lower concentration range (10–70 µg/L) than the Merck kit. In addition, the reaction vessels are equipped with a special cap to contain the arsine gas, so that exposure to the gas is limited.

Figure 1. The color scale of the Merck arsenic field kit, which is shown above, has a limited number of standards <50 µg/L. Credit: Environ. Sci. Technol.

“I think it’d be wrong to consider the test field kits as quantitative,” says Rick Johnston, a doctoral student at the University of North Carolina–Chapel Hill, who worked as a consultant for several years in Bangladesh on the arsenic problem and is currently developing ways to remove arsenic from contaminated water. “But I think they are very good at identifying the highly contaminated wells. The kits rarely have false negatives at the high concentrations,” he says.

From a public health standpoint, this is important, because a false negative could cause wells with arsenic levels >50 µg/L to be painted green or safe. WHO recommends a maximum arsenic level of 10 µg/L in drinking water, but the guideline in many developing countries, including Bangladesh, is still based on the old U.S. guideline of 50 µg/L, even though the United States now recommends a maximum of 10 µg/L. Although the percentage of wells incorrectly painted green appears to be much lower than the percentage of wells incorrectly painted red, Chakraborti and colleagues found that 7.5% of the green wells sampled in their study were actually unsafe. On the flip side, they found that nearly 50% of the wells painted red actually contain safe drinking water. Because of the scarcity of uncontaminated water in Bangladesh, mislabeling nearly half of safe wells has major socioeconomic impacts, Chakraborti says.

UNICEF is still using field kits to test wells in Bangladesh for arsenic, but it is continuously changing products, says Rosenboom. In 2003, UNICEF will begin using the recently redesigned Arsenator, originally developed by Walter Kosmus at the Karl-Franzens University of Graz, in Austria (Environ. Sci. Technol. 2001, 35, 288A). The instrument, which is being sold through the UK-based instrument company, Wagtech, uses a battery operated hand-held photometer to read the color stain, for better accuracy and precision. In addition, UNICEF is keeping an eye on a kit being developed at Mahidol University in Thailand, which also uses a hand-held photometer, and another kit being developed at Columbia University based on the molybdenum-blue method, which does not produce arsine gas.

The U.S. EPA, through its Environmental Technology Verification program, recently evaluated four portable arsenic analyzers, and two of them—Peters Engineering’s As 75 and Industrial Test Systems, Inc.’s Quick test arsenic system—performed reasonably well, says Johnston. The Peters Engineering system is based on the same technology as the Arsenator, and uses a photodiode to measure the color change over a relatively narrow concentration range. The Quick test, on the other hand, does not include photodiode detection and is another version of the mercuric bromide stain method that uses a visual color scale, with a detection limit of 5 µg/L.

A Bangladeshi woman pumps water from an untested well. Several wells in the region still need to be screened for arsenic. Credit: UNICEF-Bangladesh

Rosenboom and his colleagues at UNICEF believe that the Arsenator technology is currently “as good as it gets”, in terms of field equipment that can be used by the lay person, although they acknowledge that field kits are less accurate than laboratory based methods. As a result, UNICEF specifies that 3% of all samples must be re-tested using laboratory methods. In 2001, this meant almost 10,000 samples were tested in laboratories. In 2002, that number rose to 15,000. Any more than that is not feasible with the current laboratory capacity in Bangladesh, says Rosenboom.

Expanding testing facilities
Several laboratories in Bangladesh, including four belonging to the Department of Public Health Engineering, are being equipped with AAS instruments, but Rosenboom is not convinced that this will improve the situation. “If all wells had to be tested by laboratory methods, progress would come to a grinding halt,” he says. “UNICEF alone tested 700,000 wells in 2002. Transporting over half a million samples per year to a number of laboratories, analyzing them reliably at the trace level, and reporting back the results to the program implementer introduces too many chances for error to be comfortable with. Also, it would add years to the implementation timeline, and millions to the costs,” he explains.

Chakraborti disagrees. “Based on our analysis of around 200,000 water and biological samples from the arsenic affected areas of West Bengal and Bangladesh by FI-HG-AAS, we have proven that if thousands and thousands of samples are measured by FI-HG-AAS or similar techniques, the cost will be less compared to field kits,” he contends. The average cost of the field kits is about U.S. $2 per analysis, which includes the salary of the field workers and maintenance expenses, according to Chakraborti.

Rainwater harvesting is heavily promoted by UNICEF-Bangladesh at the household and institutional level, particularly for schools. Household tanks, such as those shown in the above photo, have a capacity of 3200 liters. Credit: UNICEF-Bangladesh

Nonetheless, even if the equipment were available, having sufficiently trained personnel to operate it remains a problem. “What is actually needed in Bangladesh is a continuing program of training and quality control to help laboratories raise the quality of their analyses,” advocates Rosenboom, who suggests starting a formal accreditation program to increase confidence in laboratory results.

Accredited laboratories in the developing world, however, are probably years away, and even with the use of field kits, it is doubtful that all organizations involved in the arsenic mitigation effort will meet their deadlines. BAMWSP worked out an agreement with the World Bank, which financed a U.S. $50 million loan for the effort, to complete screening and mitigation of wells in 147 upazilas or subdistricts in Bangladesh by June 2003. According to Rosenboom, that work had not even begun as of this past October, and BAMWSP has asked the World Bank for a second extension. It is unclear whether the World Bank will grant them the extension.

Repeat testing
Some scientists, including Chakraborti, believe the arsenic problem is getting worse. Many wells that were declared safe are now showing high levels of arsenic, he says. But other researchers say they have not seen evidence to support this kind of trend. “I think you will see a natural fluctuation on the order of 15–20%,” says Johnston, referring to seasonal and temporal changes in arsenic groundwater concentrations. The fluctuations are not always in the same direction, he says, and they are not dramatic. “It is not as if wells veer from red to green to red again,” adds Rosenboom.

Although it would be nice to repeatedly test the wells, just testing them once “is already proving to be a mammoth undertaking,” says Rosenboom. Although repeat testing will not be carried out by the government or donor agencies, “well owners are encouraged to retest their wells, preferably twice a year, and not every year at the same time,” he says. Private testing services, however, are currently not available in Bangladesh. UNICEF is working to change that and in 2003, hopes to develop new approaches to make private testing available. —BRITT E. ERICKSON