Sunday, March 23, 2014

Dealing with LOW OELs

Recently a colleague wrote to me with the following comment:  “I was gnashing my teeth today over NIOSH's ethylene oxide REL of 0.1 ppm… Would you be willing to write a blog exploring this?”

My last blog discussed the new NIOSH Guidelines on setting OELs (at least the OELs for carcinogens) on the basis of putative risk and that the quantitative risk level would be set at exposure that renders approximately a 1 in 1000 lifetime risk of getting cancer from the exposure..  This is known as a Risk Based Occupational Exposure Limit (RBOEL) and represents the top of the OEL Hierarchy that was the topic of the lead article in this month’s American Industrial Hygiene Association SYNERGIST. 

I have not done so but apparently, when you run the numbers, the putative risk of cancer during a working lifetime exposure to 0.1 ppmv ethylene oxide is about 1 in 1000 and this RBOEL will be forwarded by NIOSH as their REL. 

Well this new REL is 10 fold lower than the current OSHA 8 hour average OEL of 1 ppmv and 5 fold lower than the current OSHA action level of 0.5 ppmv.   Indeed, ethylene oxide is a gas and I am quite sure there are some uses of this gas that are well controlled to the current OSHA limits but would not be controlled under the new REL.  

Does this mean that the old OEL is dangerous?  The answer is, not necessarily.  The old OEL could be safe even though when you run the standard  model for cancer risk, the risk at 0.5 ppmv would come out to be significantly higher than 1 in 1000.   An important question remains: Is this predicted risk real?  You may remember that any level of risk at these low concentrations including the level of estimated risk at this exposure of 0.5 ppmv is putative which according to the dictionary definition of putative means it is “generally thought to be or to existeven if this may not really be true”[emphasis added].   The bottom line is that we do not have the scientific tools at this point to tell us that it is true; however, by using some assumptions it fits the above definition.   You should be aware that there remains a lot of uncertainty about the true risk at any of these exposure limits.

So how does one set any OEL, especially one that might have the strength of law?    OEL setting always has and always will boil down to a matter of politics.    On its face, if we set the OEL lower it means less risk to human health from exposure to this chemical.  Right?   Well, this is not necessarily true.   Some folks will argue that there is NO threshold for a cancer dose response especially for an alkylating agent like ethylene oxide; however, from my perspective that assertion has not been proven.   A concentration of 0.5 ppmv may be below a true threshold in humans.   I just do not think the science is there to prove otherwise.

Does that mean we should not lower the OEL given the current model and an appropriate sense of precaution?  Indeed, what if it is true?   Are we willing to allow this many cancers as a result of these workplace exposures?  The fact remains that this level of quantitative risk is what the chosen model is telling us albeit with assumptions and a considerably amount of uncertainty.   

The reality is that lowering this (and other carcinogenic OELs) appears to be where we are heading politically.  Also, it may be that epidemiological studies can and will back up the need for this change; however, the lower OELs will come at a cost.   Lower limits mean more controls and more resources dedicated to managing this putative risk.   Resources are, by definition, finite and if taken from one area have to come out of another.  It is a cost and some will probably bear the cost more than others but resources will have to be shifted and this could have untoward consequences even to health and safety.   I will not go into the details of it here but it has been fairly well demonstrated that being generally poorer as a society (less output of goods, less jobs, less taxes) could mean more risk to human health.

I am not skilled at politics or economics and I do not have the answers but I did promise to at least explore the issue.   My opinion is that if low OELs are truly problematic from an economic perspective all the elements of these changes need a holistic investigation of the entire spectrum of consequence to workers, the economy and the public health.   If overall it is deemed worthwhile using an open and democratic process, the the OELs should be lowered.

One real advantage of RBOELs for me is that one has a quantitative basis for the exposure limit and, if properly done, a description of the uncertainty associated with that limit.   For me that represents a significant increase in transparency over an OEL setting system based on safety factors born of expert judgment.
From a scientific perspective, I believe we ultimately have to get much better at understanding what really is happening at the human tissue level when workers are subjected to concentrations at current and recommended OELs.  Reduced uncertainty and more confidence will ultimately be cost-effect.  It will translate into spending money only where we really need to in order to reduce exposures and real risk while avoiding over-regulating and its inevitable waste.  

A look at a few dose-response models for carcinogens (in addition to the current linearized low-dose model) and how we might lower the uncertainty of the risk at the OEL will be the topic of a future blog.


  1. What happens if the limit set is not measurable nor quantifiable by current technology? How are we to assess the exposures? Do we wish to measure or assume we are or not exposing workers. Too often the politics of setting standard lower get in the way of reality. Cal/OSH and ACGIH, along with NIOSH are guilty of setting unrealistic values. Think about this as you ponder the issue of setting lower standards.

  2. Dear Anonymous,

    My sense is that in many if not most cases this is not an issue or the technology can be pushed to lower the LODs. Indeed, this does not present a problem in the example of ethylene oxide. When it is the case, given some good chamber work we can always model the exposure and often do as good as job in this manner as with monitoring. The point being that we should not fail to set a limit because of some current or perceived limitations related to analytical science. Where there is a will do so then a technical means will be found to do it.