Mis-lead
Water utility managers and public-health officials may be getting the wrong message about what happened during Washington, D.C.’s drinking-water crisis.
At water conferences across the U.S., scientists and engineers are talking about lessons learned from the record levels of lead in Washington, D.C., drinking water that caused panic in the city in 2004. Government agencies describe the lack of harm from the incident; this has prompted many water and public-health professionals to argue that the D.C. experience shows that lead in drinking water is not a health threat. As a result, some experts now question the need for complex and costly technologies to control corrosion and keep lead levels low.
But are the water experts being misled? An extensive 2-year investigation by Virginia Polytechnic and State University corrosion engineer Marc Edwards, who initially identified the severity of the D.C. problem, and further ES&T reporting reveal that the federal and local agencies charged with overseeing the D.C. water system used flawed science to try to quiet public concerns.
Lessons in lead
In March 2004, John Morrow, director of the Public Health Department in Pitt County, N.C., had a mystery on his hands. His office had identified a 1-year-old child in Greenville with elevated blood lead levels, but extensive investigations in the boy’s home and the nearby environment failed to turn up a contamination source. Then, a water sample taken as part of an unrelated investigation revealed high levels of lead in the family’s drinking water. As the child’s blood lead level climbed above 20 micrograms/deciliter (µg/dL)—two times the Centers for Disease Control and Prevention (CDC) level of concern for children under 6 years old—Morrow turned to experts in D.C. for advice.
Just 3 months earlier, D.C. residents had learned from a Washington Post story that hazardous levels of lead had been present in their drinking water for several years. The problem was caused by a switch from free chlorine to chloramine disinfectant in order to meet new U.S. EPA regulations. D.C. officials pointed to old lead pipes called service lines, which bring water to homes, as the source of the lead.
Like D.C., Greenville’s water provider, Greenville Utilities, had switched from free chlorine to chloramines for disinfection in 2002. The switch seemed to be related to the elevated lead levels. But none of Greenville’s 565 miles of water pipes contained lead.
Faced with what appeared to be a similar problem, Morrow found materials on the web from Tee Guidotti, health adviser to D.C. Water and Sewer Authority (WASA), the capital’s water provider, and director of occupational medicine and toxicology at George Washington University. Guidotti downplayed the role of water. “These all indicated to me that drinking water lead and blood lead are not correlated,” says Morrow.
Indeed, on March 31, 2004, D.C. Department of Health (DOH) interim chief medical officer Daniel Lucey told the Washington Post, “We are not seeing any widespread lead toxicity attributable to the water in D.C.” Lucey’s comments were prompted by a preliminary CDC study that concluded blood lead levels had not risen appreciably as a result of D.C.’s tainted water, even in homes where concentrations were unusually high—300 parts per billion (ppb) or more. Meanwhile, sampling data from local schools and day-care facilities suggested that drinking-water lead concentrations in these places were not extraordinarily high. On the basis of these findings, many accepted the statements of agency experts, that the “D.C. lead crisis” was much ado about nothing and that “drinking water is at most a minor source of lead for children” (The Washington Post, May 9, 2004, p B1).
But a careful investigation of the D.C. studies gives a different picture. The crucial assumption that the lead service lines were the only major source of the toxic metal, along with the surveys for lead in blood and in drinking water at D.C. schools and apartments, is deeply flawed and misleading, say experts familiar with the work.
The change in disinfectant did cause mineral scales inside D.C. lead service lines to dissolve, says corrosion expert Michael Schock with EPA’s Office of Research and Development. However, the more-corrosive water also eroded lead solder, which sent particles down the pipes, and leached the metal from brass plumbing in homes. He points out that brass faucet bodies and necks, shutoff valves, water meters, and other plumbing components usually contain lead, even if the brass is labeled lead-free. That is because Congress defined in the 1996 Safe Drinking Water Act amendments that “lead-free” plumbing pipes and components could have up to 8% lead.
Edwards can document at least two children in one D.C. family whose high blood levels are due to exposure from drinking water [web (see graph below). And in Greenville, particles of lead solder were eventually identified as the source of the contamination for the 1-year-old boy and another child with elevated blood lead levels.
However, misconceptions from the D.C. lead crisis persist, even as other water utilities switch from chlorine to chloramines disinfection to comply with EPA’s Stage 2 Disinfectants and Disinfection Byproducts Rule. Experts warn that problems with lead in water—and perhaps further cases of elevated blood lead levels—could be avoided if scientists, regulators, engineers, and public-health officials fully realized the true lessons of the D.C. lead crisis.
Lead service lines are the source
Early on, officials in D.C. decided to focus on lead service pipes as the only significant source of contamination. Through numerous Freedom of Information Act requests, Edwards has pieced together a partial chronology of the discussions that focused attention on the lead service lines. Despite repeated requests, officials at DOH and WASA declined to speak to ES&T for this article.
From the start, some at DOH expressed the hope that the data supported a focus only on lead service lines, because this would limit the problem. Of the about 123,000 residences in D.C., only about 23,000 are estimated to have lead service lines.
On February 8, 2004, James Collier, DOH’s water-quality division director, emailed his colleagues: “I believe that if we can get away from old lead-soldered copper plumbing and isolate on the lead service lines, we can calm down all of the public except for the 23,000. If we stay with the old copper lead-solder plumbing, we have 123,000. Just make sure that the data supports this because we can not afford to be very wrong.”
A day later, WASA mailed an information letter on the lead crisis, which emphasized the problem of lead service lines, briefly mentioned solder, and ignored brass plumbing components. But within the agencies involved, a very different discussion was taking place.
On Feburary 12, EPA Region environmental scientist Lisa Donahue emailed agency colleagues: “We continue to miss a ‘teachable moment’ by maintaining the emphasis on the service lines as a perceived sole source of lead contamination. The first draw [sic] samples that put WASA into this situation came from plumbing containing brass faucets, fixtures, and valves. Until the water has been tested below 15 ppb, shouldn’t those high risk populations be particularly cautious? If the water is corrosive, the brass will leach lead.”
At that time, Donahue suggested that future public information letters warn pregnant women and children in all homes that tested above 15 ppb—EPA’s action level—not to drink the water. This would have expanded the city’s focus to homes without lead service lines.
Donahue’s language was never accepted. And throughout the crisis, DOH focused its public-health intervention efforts almost exclusively on those homes with lead service lines.
Schock was the corrosion expert on the technical evaluation work group assembled to advise authorities on the crisis. In response to questions about the source of the lead, he replied on March 22, 2004, “What I find hard to believe is why, with the known contamination potential from leaded brasses currently being certified by NSF [for plumbing standards] and sold, plus some old solder around, does there seem to be so much effort to just focus on the service lines?”
Richard Rogers, chief of EPA’s Region 3 drinking-water branch, replied, “This is being driven as much by public relations and politicians as by what makes sense most other ways.”
Indeed, later data published by WASA showed that 15% of the homes with very high (above 150 ppb) lead did not have lead service lines. But only residents with lead pipes received the special health warnings, bottled water, and filters from WASA.
D.C. schools and apartments
On April 29, DOH and WASA released the results of sampling for lead in D.C. public schools. Only 4% of the 1976 water samples tested from 130 schools had lead levels above 20 ppb, the action level specified by the 1988 Lead Contamination Control Act, which covers drinking water in schools and day-care centers.
Given the high levels of lead in drinking water at the time, experts say that these results were surprisingly low and had the effect of reassuring people. But there are strong reasons to think that the sampling strategy was flawed. For instance, a neighboring utility that used the same water source as D.C., but which has extremely good lead-corrosion control, did not have a single home sample above the EPA action limit. Yet 23.5% of the water taps in schools served by this utility were more than 20 ppb, says Edwards.
The school sampling was ordered by EPA Region 3 administrator Donald Welsh, who added that “Sampling must follow EPA protocol.” The agency’s protocol for sampling schools calls for first-draw samples to be taken after the water has been in contact with pipes for 8–12 hours. But a very different protocol was followed in D.C. schools. The day before sampling, staff were directed to remove the aerators and flush all of the drinking-water lines in the entire building from top to bottom. After letting the water sit overnight, they were directed to slowly fill sample bottles.
“If I did not want to find a lead problem where a serious problem existed, this is the protocol I would write,” Edwards says. Because these schools do not have lead service lines, the principal source of lead leaching would be from solder or brass. EPA’s own literature notes that particles trapped in the aerators serve as a key source of lead and that flushing sediments from the lines can reduce lead exposure. Collecting the water slowly minimizes the chance of mobilizing lead particles.
Edwards has followed the standard EPA protocol and the modified one used in D.C. schools to sample for lead in buildings that do not have lead service lines. The modified approach decreases lead levels in faucets with aerators by 2–3 times the values from the EPA protocol; in some situations the decrease is as much as 200 times. Edwards and his students have been unable to get permission to collect samples in D.C. schools using the standard EPA protocol.
A consulting engineer familiar with sampling in schools agrees with Edwards that the D.C. procedure was odd. “They flushed the devil out of those schools,” the expert says. A Region 3 EPA official who spoke on condition of anonymity tells ES&T that the sampling was not meant to reflect what children were being exposed to at the time. “That study was designed to look at exposure in the future, not exposure in the past.”
Another scientist sums up these criticisms: “The survey of lead in D.C. school drinking water was unlikely to reveal both the actual previous situation to which children and teachers were really exposed, not to mention the worst-case scenario, which is the intention of the law.” The same modified protocol was eventually used by WASA to assess problems with lead in D.C. apartments—and found relatively low levels of lead. This observation was used to justify WASA and DOH’s exclusive focus on homes with lead service lines.
CDC’s survey
When asked to list some of the most important lessons from the D.C. lead crisis, Guidotti says that “Drinking water is a minor contributor to lead exposure, but takes on huge significance because people become worried when they hear that drinking water, which they depend upon, may be contaminated.” Guidotti has emphasized this point many times when he has spoken at scientific meetings throughout the country (see PowerPoint presentation).
Everyone agrees that any exposure to lead is detrimental to children, and a growing number of studies quantitatively assess the effect of low-level exposure to lead, says John Rosen, a pediatrician and national expert on lead poisoning at Montefiore Medical Center. Last year, a study by Bruce Lanphear at Cincinnati Children’s Hospital Medical Center and colleagues associated a drop of almost 4 IQ points with blood lead increase from 2.4 to 10 µg [already defined on first use for web] of lead per dL of blood (Environ. Health Perspect. 2005, 113, 894–899). This study is consistent with other reports, say experts.
In the midst of the lead crisis, EPA’s Office of Water asked the agency’s National Center for Environmental Assessment (NCEA) to evaluate the effects of lead in drinking water on children’s blood lead levels with an exposure model. In March 2004, NCEA delivered its assessment [130KB PDF]: The blood lead levels of infants up to 1 year in age who drink formula made with tap water “are sensitive to drinking water lead concentrations.” The model predicted that infants’ blood lead levels would increase to approximately 6 µg/dL for a water lead concentration of 50 ppb, for 11 µg/dL at 100 ppb, and to a dangerous 20.8 µg/dL for a concentration of 200 ppb.
However, Guidotti, Wasa and DC DOH have frequently noted in presentations that when CDC measured blood lead levels in the residents of approximately 98 homes with drinking-water lead of more than 300 ppb, the study did not find elevated blood lead levels.
But Mary Jean Brown, head of CDC’s lead poisoning prevention branch and the principal author of the study, doesn’t agree with this analysis. She tells ES&T that up to a year separates collection of the water samples and the blood samples. “This study does not say that 300 ppb lead in drinking water is safe,” says Brown.
As Edwards points out, many of those tested by CDC had been notified that their water contained lead at more than 20 times the EPA action limit months before their blood was drawn. It is likely that many began drinking bottled water or using water filters. Since the half-life for lead in the blood is about a month, this was more than enough time for blood lead levels to drop, he adds.
When asked by ES&T, Guidotti agrees that the CDC study is not conclusive. “This was an ecological study, and ecological studies are weak at proving associations,” he admits. “All of D.C. was intensely aware of the lead problem—not just through newspaper reports but through public meetings and announcements in churches. People rapidly started using filters or bottled water,” he says. “It is a major misinterpretation of the data to say that this study shows that 300 ppb in drinking water is not associated with an increase in blood lead levels.”
Brown tells ES&T that she intends to look into the issue of the study’s interpretation and seek ways to clarify its significance. “If misinterpretation is widespread, we’ll have to do something, because that’s not what this study is saying.”
Brown may have to work quickly. Catherine Karr, director of the Northwest Pediatric Environmental Health Specialty Unit at the University of Washington, has recently been investigating lead in Seattle schools’ drinking water. According to Karr, CDC’s study is “very influential because it is one of the only studies that presents data on water lead and blood lead.” She was surprised to learn from ES&T about the study’s limitations. “The study is misleading,” she says. “They could have made the sampling gap [between exposure and measurement] clear by just adding one small sentence saying that there was a sampling gap. Why didn’t they do that?”
“At a minimum, it seems easy to understand how otherwise responsible public-health officials believed that the takeaway lesson of Washington, D.C., is that more than 300 ppb lead in drinking water did not significantly elevate blood lead or otherwise harm the public,” says Edwards. “It may take years to correct this mistaken belief,” he adds.
Russian roulette
The Greenville incident illustrates that under some circumstances, the corrosivity of drinking water can be altered so that it aggressively attacks lead solder. Pieces of solder from the water pipes can then detach sporadically and contaminate the water. This is difficult to monitor because multiple samples can be collected from a tap with relatively low lead, but occasionally a sample can be collected that contains more than 15,000 ppb—as much lead as could be consumed in paint chips. Edwards terms this disconcerting problem “Russian roulette” and says that it explains sampling data collected by EPA and his own research group in D.C., Greenville, and other locations.
“I would like to know how common it is for lead in drinking water to elevate blood lead levels,” says Morrow. “We’ve tried to get parents to bring in their kids. We’ve tried to get doctors to test all 1- and 2-year-olds. But we’ve only tested about 45% of the kids, so we just don’t know.”
Neither Edwards nor any of the experts contacted for this story claim that drinking water is the major source of lead for children nationally. Public-health experts are much more concerned about chips and dust from leaded paint, says Brown. But Edwards has assembled enough evidence to indicate that lead in water can sometimes be a key source of elevated lead in children’s blood. And that may be the real lesson of the D.C. water crisis.
