Strength of Evidence
My observations from a review of the history of PERC risk assessment follow:
The NAS conclusions blow the doors off the 100 ppm OSHA
PEL for PERC. That PEL is a license to kill. I was not aware of
the strength of the evidence until I started reviewing the two
recent documents, especially the low levels of exposure that
caused cancer in the animal studies. As a reviewer of the 1986
NTP inhalation bioassay of PERC, I had proposed the motion
that identified leukemia in rats as “clear” evidence for carcinogenicity. (The wording proposed by NTP staff was that the rat
leukemia represented “some” evidence. You can read the discussion at http://ntp.niehs.nih.gov/ntp/htdocs/LT_rpts/tr311.pdf).
The estimated cancer risk at 1 ppm is far more than the canonical 1-in-1,000 reference for “significant risk” derived from the
1980 United States Supreme Court ruling vacating OSHA’s
then-standard for benzene. Even skipping the extrapolation,
risk increases were found at the PEL. The least protective EPA
RfC is 1,000-fold below the OSHA PEL and 250-fold below the
ACGIH TLV, taking into account the 8-hour (occupational) versus
24-hour (RfC) daily exposure. The 25 ppm TLV also leaves
workers at risk.
The volume of work in these reviews staggers the mind. The
EPA draft risk assessment is 550 pages and is based on multiple
detailed calculations as well as extensive literature review. I
couldn’t count the references conveniently. The NAS report was
140 pages, including 12 pages of references. It occupied nearly
two dozen respected scientists over three years. The IARC
monograph, a lightweight at 63 pages, included 17 pages of
references.
The EPA document isn’t a complete risk assessment. Even
though I’ve referred to it as a risk assessment, it’s actually a
“hazard identification” followed by an “exposure-response assessment.” A canonical risk assessment also includes exposure
assessment followed by a risk characterization, essentially multiplying risk rate by exposure levels to get a body count. However, a regulation could be based on a perception of an
unacceptable rate of risk for any exposed population, without
detailed compilation of exposure levels. ACGIH sets TLVs without a quantitative risk characterization—but also without any
explicit reference to an acceptable risk.
It’s time to introduce RfCs and benchmark doses into IH
practice. RfCs are applied to non-cancer effects assumed to
have a threshold. EPA selects a target health effect and the
studies showing the level of exposure or dose associated with
that effect. For PERC, that effect was neurotoxicity rather than
reproductive, liver and kidney toxicity. The point of departure
for the evaluation may be the No Observed Adverse Effect Level
(NOAEL) or its statistical equivalent, a benchmark dose. The
NOAEL is therefore a dose at which 10 percent of the population suffers the adverse effect. The risk rate at the NOAEL is
100-fold above the Supreme Court’s 1-in-1,000 reference for
significant risk. Generally, uncertainty factors are applied: 10-
fold each for variability in the human population, subchronic to
chronic exposure, and animal to man, with an additional factor
of 10 for the Lowest Observed Adverse Effect Level (LOAEL) to
INSIGHT | DEPARTMENT
NOAEL and database uncertainty. For PERC, EPA applied a factor of 300 to neurological effects in studies of people.
The IRIS system is a result of the general perception that different programs within EPA—air, water, toxic substances—
should rely on the same potency values. This was the opinion
of the 1983 NAS Risk Assessment Committee on which I served.
However, the committee also believed that it was improper to
centralize full risk assessments because considerations—
especially exposure considerations—were medium-specific. As far as
I can tell, IRIS and its development process, inference rules and
review procedures are policies implemented by EPA rather than
regulations or implementation of legislation.
No exposure scenario has been regulated based on all this
work. More work will occur and more time will pass as EPA responds to the NAS report. After that, the potency assessment
will have to be factored into specific regulations. The justification for those regulations will probably double the volume of
paper at EPA alone. This work could support OSHA rulemaking,
but nothing is in prospect.
Lessons for Controlling PERC
Some progressives will say that this scientific detail is a waste
of time, forget about a PEL, PERC should just be banned from
dry cleaning—in fact, forget about PELs altogether; let’s do substitutions and control banding. But “bans” are going to be
phase-outs at best, as small businesses are given time to adjust
(while they keep exposing workers and communities). How are
we going to evaluate the interim exposures? Control technology
directed by control banding also has to be defined and evaluated based on expected exposure reductions, which come back
to a reference level. Regarding substitution, remember, PERC
was a less toxic alternative to carbon tetrachloride and a less
flammable alternative to Stoddard solvent. That said, the NAS
report is an authoritative statement that gives the industrial hygiene community a foothold in controlling PERC.
FranklinMirer,PhD,CIH,isaprofessorintheEnvironmentalandOccupational
Health Sciences Urban Public Health Program at Hunter College School of Health
Sciences in New York. He can be reached at (212) 481-7651 or
fmirer@hunter.cuny.edu.
Resources
Environmental Protection Agency: “IRIS Toxicological
Review of Tetrachloroethylene.” [Online] Available at
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=1924
23 (2008).
National Academy of Sciences: Review of the Environmental Protection Agency’s Draft IRIS Assessment of
Tetrachloroethylene. [Online] Available at www.nap.edu/
catalog.php?record_id=12863 (2010).
