Questions remain about MCHM screening level
CHARLESTON, W.Va. -- On the evening of Jan. 9, Larry Cseh was watching TV with his wife, his daughter and his grandkids. Then, shortly before 8 p.m., the call came in from his office.
There'd been a chemical leak in West Virginia. Thousands of gallons of something called Crude MCHM made it into the Elk River. The drinking-water source for 300,000 residents was contaminated.
Public health officials in West Virginia wanted help. What could the chemical do to people? How much of it was safe to allow in drinking water?
Cseh, a senior environmental health and emergency response officer the U.S. Public Health Service, went to work. He got on the computer, went to the National Library of Medicine's TOXNET database, and looked for information on Crude MCHM.
"There was nothing in it. It was empty," Cseh recalled last week. "I was kind of panicking there."
Then Cseh saw something he could use. There was a number called an LD50 on the material safety data sheet from the chemical's manufacturer, Eastman Chemical Co. The LD50 gave an estimate of how much Crude MCHM killed half of the lab rats exposed in an Eastman study.
It wasn't much information, and it couldn't be directly extrapolated to humans and polluted drinking water. But it was all Cseh had, and West Virginia officials were desperate for help. So Cseh did some division. By 9 p.m. -- barely an hour after that initial call about the leak -- he'd settled on what's become practically a magic number in the region's water crisis: An emergency "screening level" of 1 part per million.
"Last night we discussed a screening level of 1 ppm we continue to believe will be protective of public health," Cseh told state Bureau for Public Health officials in an email message at 7:09 a.m. on Jan. 10, the day after the leak.
In the month since the Freedom Industries' chemical leak, Cseh's number has been cited over and over by state and federal officials who say West Virginia American Water supplies are fine for residents to use for drinking, bathing, or anything else they want, now that the MCHM levels have dropped below that 1-part-per-million number.
"Science says the water is safe," Gov. Earl Ray Tomblin said at a high-profile press conference last week.
But outside experts on public health and toxic chemical issues have questioned the 1 part-per-million number. They say federal officials had nowhere near enough information to really develop a health screening level, and complain that the government's number did not provide an adequate margin of safety given the lack of data about Crude MCHM's possible impacts.
And in interviews, email messages and public records, federal public health officials have continued to offer confusing statements about how and when they developed the screening level, and what data they used in their calculations.
"This episode should serve as a case study in how not to handle both public health and public trust in the aftermath of a chemical spill," said Richard Denison, an Environmental Defense Fund senior scientist who has followed the situation closely and written about it on his group's blog.
. . . does not present a health risk to customers'
In the hours after the Crude MCHM leak at Freedom Industries was discovered, state and local officials responded as they frequently do when there's a leak or leak at one of the Kanawha Valley's chemical plants.
Emergency officials and environmental inspectors said they had no reason to think the incident was that big of a deal. West Virginia American Water assured the media the leak "does not present a health risk to customers."
What quickly emerged, though, was that public health officials actually knew next to nothing about Crude MCHM's toxicity. Like thousands of other chemicals on the market, it hadn't been studied much, and the few studies that were done by the company that made it, were never peer-reviewed, and weren't publicly available in scientific journals.
For some period of time -- it's not clear how long -- officials from the state Department of Environmental Protection to the U.S. Centers for Disease Control -- were working basically from only one document, a nearly decade-old "material safety data sheet," or MSDS, from Eastman Chemical, which made the Crude MCHM that Freedom leaked into the Elk River.
DEP officials relied on the MSDS when they told the public the chemical wasn't "hazardous," a narrow definition based really only on the lack of a U.S. Department of Transportation regulation governing its shipments. The MSDS, though, also noted the chemical was hazardous under U.S. Occupational Safety and Health Administration standards, and warned exposure causes skin and eye irritation and that, "at elevated temperatures, vapor may cause irritation of eyes and respiratory tract."
It was the Eastman MSDS that Cseh would also use to set the initial chemical screening value that has been cited as the figure developed by the CDC, which is a sister agency to the Agency for Toxic Substances and Disease Registry, or ATSDR, where Cseh is actually assigned.
'An abundance of caution'
The 1-ppm number first surfaced publicly during a press conference late on the afternoon of Jan. 10. National Guard Gen. James Hoyer mentioned it almost as a passing remark, and state officials offered no immediate explanation.
Later that night, Bureau for Public Health Commissioner Letitia Tierney initially refused to answer questions about the screening level, but later provided an explanation in an email forwarded by an agency publicist.
In her email, Tierney explained that the federal officials had found only one study, the Eastman one from the early 1990s that resulted in the LD50 figure listed on the company's MSDS sheet.
Tierney wrote that the Eastman LD50 study was completed in 1990. It was actually completed in 1998.
The MSDS sheet listed the LD50 for Crude MCHM as 825 milligrams per kilogram. That means, when tested on rats, an 825 milligram dose per kilogram of body weight was enough to kill half the rats.
"The experts then took this number and calculated the uncertainty factors," Tierney wrote. "In this situation, there were two. The first uncertainty factor was translating these results from rats to humans. The second uncertainty factor took into account sensitive populations. This includes the elderly, the sick, immuno-compromised and children, amongst others.
"Uncertainty factors range from 5 to 10 percent," she wrote. "Given the dearth of data and an abundance of caution, both uncertainty factors were rated at 10 percent."
Basically, what this means is federal officials divided 825 milligrams per kilogram by 10, and then divided that result by 10 again, ending up with a figure of 8.25 milligrams of chemical per kilogram of body weight.
Because they were using an LD50 -- which considers only death, and not other non-lethal health effects -- federal officials cut that 8.25 number down to 1 milligram per kilogram, or 1 part per million, Tierney said in her email.
At the time, Tierney did not respond to follow-up questions. She suggested a reporter talk with Cseh, who did not respond to phone calls or email messages until last week, when a CDC public affairs officer set up a telephone interview with the Sunday Gazette-Mail.
Over the weekend of Jan. 11-12, CDC press officer Bernadette Burden provided an email explanation similar to Tierney's. Burden said her agency had no one available to answer questions about the West Virginia situation. Burden referred questions back to Tierney's agency.
"I'm aware of the circumstances and that's why we've provided the background that we have, but I cannot guarantee rapid response given the limitations," Burden said in an email message on the evening of Jan. 11.
Burden wrote that the 1 ppm number was "extremely conservative and protective of public health."
The CDC did not make anyone available for an interview with West Virginia media outlets until Jan. 16, the day after a Gazette-Mail reporter called CDC director Dr. Thomas Frieden at home, following days of silence by the federal agency. That was also three days after state officials and the water company began clearing residents to use the water again.
In the meantime, though, outside experts expressed serious concerns with the CDC number.
Denison, whose group works for tougher regulation of toxic chemicals, blogged that the federal agency's figure had a "questionable basis" and that West Virginia officials were trusting "shaky science."
Among other concerns, Denison said that LD50 numbers for the lethality of a chemical should not be used to develop this sort of screening level, and the formula used did not take into account potential effects of longer-term, chronic exposure to the chemicals.
"Now let me be clear," Denison wrote. "I am not saying that the level of 1 ppm is unsafe. I am saying that we have no way of knowing whether it is safe. The data needed to make that assessment simply do not exist for this chemical."
'No observed adverse effects'
When the CDC finally started talking to the media and the public, the agency's explanation of how it arrived at the 1-ppm number had changed.
Starting with interviews, a media briefing, and a website posting, the CDC said it had used a different starting point than the LD50 number Cseh initially relied upon.
Dr. Vikas Kapil, chief medical officer for CDC's National Center for Environmental Health, told reporters Jan. 16 the agency used what's called a "no observable adverse effects level," or a NOAEL. That figure, 100 milligrams per kilogram, came from a 1990 Eastman study of 4-MCHM, the main component of Crude MCHM.
Using it, CDC officials again applied three uncertainty factors -- one for variation between animals and humans, one for vulnerable populations, and a third to try to account for how little was known about Crude MCHM's potential toxicity. They came up with the same end result: 1 part per million.
The CDC's use of the LD50 study for such an important screening level -- one impacting 300,000 people -- had been questioned in the public health community. The use instead of the NOAEL figure would deflect at least some of that criticism.
What remains unclear, though, is when exactly the CDC and other federal agencies or state officials obtained the 1990 Eastman study where they got the NOAEL number.
While the NOAEL study was older than the LD50 study -- it was from 1990 -- the NOAEL study was not listed on Eastman Chemical MSDS sheets published in 2005 and 2011.
In an interview, Cseh said he obtained the information sometime after 9:30 p.m. on Friday, Jan. 10, and didn't finish reviewing the material until the following day.
But Burden, the CDC press officer, was still talking only about the LD50 study in an email message late Saturday afternoon. And as late as early in the morning on Sunday, Jan. 12, Cseh sent an email message to state officials that still referenced only the LD50 number's use in developing the screening level, according to records obtained under the Freedom of Information Act.
Barbara Reynolds, another CDC press officer, said the initial recommendation of 1 ppm was based on "the limited science available" on the night of Jan. 9.
"The next day, this recommendation was validated by additional scientific studies including" the NOAEL, Reynolds said in a written response to Gazette-Mail questions. "A subsequent review by a Federal Interagency workgroup further validated the 1 ppm recommendation."
Also unclear, though, is exactly when and how the CDC obtained the Eastman studies.
Cseh said in his interview that he obtained them late on the night of Friday, Jan. 10. But at the time, he said, they were still considered proprietary work from Eastman, and the CDC could not make them public. In her email to the Gazette-Mail, Tierney also said that the Eastman study "is not publicly available."
"It is Eastman proprietary information," Tierney said.
But in an email Sunday, Jan. 12, when asked about any studies the CDC had obtained from Eastman, agency spokeswoman Burden said, "Eastman provided no studies or information to our agency. The information that is referenced in the response we sent you including studies that have been conducted in the past is available in the literature in the public domain."
Eastman posted its MCHM studies on its website -- releasing them publicly for the first time -- on the evening of Jan. 16, hours after Rep. Henry Waxman, D-Calif., wrote to the company asking for the studies to be released.
A 'clearly daunting task'
Even once the CDC started using the NOAEL number in their screening level calculations, experts from the public health community continued to criticize what they had done.
In a post for her group's blog, Natural Resources Defense Council senior scientist Jennifer Sass argued that the agency could have done the math differently, and come up with a more protective screening level.
Sass explained that the Eastman study used four chemical doses: No chemical (control), 25, 100, and 400 milligrams per kilogram body weight per day fed to each rat. Eastman's authors said the mid-dose of 100 milligrams per kilogram had no effect, and therefore should be used to calculate a "safe" level for people.
"But that's not what the data shows," Sass wrote. Male rats had body weight decreases at all dose levels, including the low dose of 25 and the mid dose of 100, she wrote.
"The effect showed a dose-response trend, increasing with increasing dose, but the trend was not statistically significant," Sass wrote. In female rats, the low- and mid-dose groups weighed 'slightly more than controls' but no specific measurements were reported and we are to trust the study authors that it was not statistically significant."
Sass wrote, "It is hard to get statistical significance in studies like this where so few rats are tested at each dose, given the natural variation between individual animals.
"In statistical-ese, we say that the study is underpowered -- its ability to detect an effect is very poor because the sample size is so small," Sass wrote. "So, when such a study actually finds an effect, like this body weight decrease, despite its lack of power, it suggests a treatment-related effect.
"The chemical company may 'overlook' the effect at low doses, but government scientists charged with protecting the public's safety shouldn't be so dismissive of the toxic effects seen at the mid- and low-dose."
Sass said if the CDC had used the low-dose figures from the Eastman study, the agency would have come up with a screening level far lower - 0.025 parts per million, or about 40 times more protective than the 1 ppm level.
Denison, the EDF scientist, said state and federal officials "were faced with the clearly daunting task of having exceedingly limited information to go on as they tried to assess and communicate about the risks of a chemical contaminating the tap water of hundreds of thousands of people.
"But instead of forthrightly acknowledging the enormous uncertainties, they instead rushed to establish a 'safe' level using that inadequate information, took many days to come clean about how they had derived the level, repeatedly claimed erroneously that the level provided a large safety margin, and then as new uncertainties arose -- more than one chemical actually leaked, exposure by inhalation during bathing had not been considered, the distinctive odor of the chemical was still being detected in homes and schools even after following the correct flushing procedure -- they continued to claim they had taken into account all contingencies," Denison said. "The result is a situation where, a month after the spill, there are still far more questions than there are answers."
Reach Ken Ward Jr. at firstname.lastname@example.org or 304-348-1702.