RfD and Consensus...Part 1

I recently added a section called "consensus" to the Wikipedia page on "Reference Dose."

Yeah, that's what I do for fun.

I was looking for some supporting information to help me put my AHMP presentation together and came across an EPA memo titled:

WASHINGTON STATE DEPARTMENT OF HEALTH’S EVALUATION OF EPA’S PERCHLORATE DRINKING WATER PRELIMINARY REMEDIATION GOAL (PRG).

Back to the question of "safe" is it?  Here is what the memo says about perchlorates found in the Deep Creek area in the state of Washington:

There are no federal or Washington State drinking water standards for perchlorate. However, EPA uses a 24.5 ug/l perchlorate drinking water equivalent level (DWEL) as a preliminary remediation goal (PRG). 

In looking at how perchlorates enter into the human receptor, EPA has decided that the predominate way would be through drinking Deep Creek water which has perchlorate contamination.

Though there is no MCL - "maximum contaminant level" - for perchlorates, the EPA uses the DWEL - "drinking water equivalent level" to establish a "safe level" of perchlorates in the water that may be consumed.

The DWEL is the basis for establishing the MCL, so this is a reasonable approach   The DWEL is calculated for non-carcinogens and uses the reference dose (RfD).

The DWEL, therefore, is defined by EPA as:

"a drinking water lifetime exposure level, assuming 100% exposure from that medium, at which adverse, noncarcinogenic health effects would not be expected to occur."

When the EPA establishes the "safe" level of a contaminant in drinking water it bases it on the "lifetime health advisory" derived as follows:

The Lifetime HA for the drinking water contaminant is calculated from its associated Drinking Water Equivalent Level (DWEL), obtained from its RfD, and incorporates a drinking water Relative Source Contribution (RSC) factor of contaminant-specific data or a default of 20% of total exposure from all sources. 

As you can see, the RfD is the foundation upon which all these "safe" levels are derived.

Where we have confusion is when we compare a concentration in one media - such as apple juice - with the regulatory level established for drinking water.  Without an understanding of the RfD for non-carcinogens and risk factors for carcinogens, comparing what is found in the apple juice to what is acceptable in drinking water will make the apple juice appear to be tainted.

Case in point.  From the Dr. Oz Website: 

In 2006, the Environmental Protection Agency (EPA) enacted a stringent health standard with respect to arsenic levels in drinking water, stating concentrations should not exceed 10 parts per billion (ppb). Interestingly, the actual goal for arsenic exposure from drinking water, what is termed the maximum contaminant level goal (MCLG), is zero ppb. That bears repeating – the goal is no arsenic in our drinking water. The 10 ppb level is simply as close as we can reasonably get considering our natural exposure to arsenic in the environment and other limitations. At that level, almost all experts agree our drinking water is quite safe.

So if our drinking water is "quite safe" why say this right after?

Shouldn't the same goal or, at the very least, similar science-based exposure guidelines, be in place for the juices we commonly give to our children?

It is, in my opinion, due to a disconnect in how a "safe" level is derived, a disconnect on how the the RfD is calculated, and a disconnect in how a "goal" and "risk" are different.

Let's look at benzene, for example. 

Source

Now benzene is kind of an odd bird, which makes it a good example on how this works.

Because benzene is considered a human carcinogen, there may be some degree of carcinogenic risk even below the MCL. Based upon EPA calculations, the EHP estimates that drinking water containing 5 ppb benzene would be associated with an increased lifetime risk of cancer in the range of between two and eight in one million (2 to 8 excess cancer cases in 1,000,000 people exposed). This estimate is based on a daily intake of two liters of water per day for 70 years. (1)

For this reason - carcinogenicity - benzene is given an MCLG of "0":

MCLG: Maximum Contaminant Level Goal. A non-enforceable health benchmark goal which is set at a level at which no known or anticipated adverse effect on the health of persons is expected to occur and which allows an adequate margin of safety. (2)

The EPA has also developed a toxicity value (Reference Dose or RfD) for non-cancer effects based on a human study.

A decrease in cells which are vital to immune system function (lymphocytes) was the most sensitive effect of several measured blood parameters. The adult drinking water equivalent level (DWEL) for this RfD is a benzene concentration in water of 140 ppb - 0.14 mg/L - ppm.

The benzene DWEL, which is the "lifetime exposure level assuming 100% exposure from that medium," is 0.1 mg/L.  This is based on the following formula:

Source

Plugging the RfD into the formula, we get:

DWEL (mg/L) = (0.001 x 70) / 2 = 0.14 mg/L = 0.1 mg/L or "ppm"

Because we know that there is a risk of cancer, the MCLG is set at zero.  Because we know that zero is unattainable, and we know that there is also a non-cancer risk, we can set the legally enforceable MCL for benzene at 0.005 mg/L or 5 ppb.

When Dr. Oz posts:

Interestingly, the actual goal for arsenic exposure from drinking water, what is termed the maximum contaminant level goal (MCLG), is zero ppb. That bears repeating – the goal is no arsenic in our drinking water. The 10 ppb level is simply as close as we can reasonably get considering our natural exposure to arsenic in the environment and other limitations. At that level, almost all experts agree our drinking water is quite safe.

He is acknowledging that the current standard for arsenic is safe even though the goal is zero.  This is the same with benzene and all the other carcinogens.  We want zero, it is safe at the MCL.  Why is it safe?  Because the MCL and MCLG is set lower than the DWEL:

The MCLG is then derived by considering other known or potential sources of exposure, using the relative source contribution (RSC) factor.

• MCLG (mg/L) = DWEL x RSC

The RSC from drinking water is based on actual exposure data, or, if data are not available, a value of 20% is assumed for effects based on lifetime exposure. This allows 80% of the total exposure to come from sources other than drinking water, such as exposure from food, inhalation, or dermal contact. For the few MCLGs based on adverse effects related to exposure in children, an RSC of 100% was usually applied because the source of exposure for the critical study was drinking water. However, in more recent assessments, even when actual data from other sources are available, EPA uses a maximum RSC value of 80% to allow for potential unidentified sources. (3)

In other words, for a chemical such as benzene, a non-cancer RfD is established that generates a DWEL of 0.14 mg/L.  That DWEL is then multiplied by about 0.8 (80%) to compensate for all the other ways the chemical gets into the receptor.  For benzene, however, the MCLG is set at zero and the MCL - the value set as close to the MCLG as feasible using the best available analytical and treatment technologies and taking cost into consideration - is set at 0.005 mg/L.

• We consider the DWEL safe because it is based on the RfD.
• We consider the RfD acceptable because it is based on "an estimate (with uncertainty spanning perhaps an order of magnitude) of a daily oral exposure to the human population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a lifetime." (3)
• The MCL is less than the DWEL
• The MCLG can be less than the MCL and is set at "zero" for carcinogens.

How low can you go (/queue music)

All the way to zero.

Which brings me back to my original question of the RfD.  If the RfD is the foundation, how confident are we in the number we are using?  As the EPA writes:

A change in the RfD could lead to a change in the MCLG and thus possibly also in the MCL. (3)

A good read on all of this can be found here.

Next post: RfD and Consensus...Part 2


.