Short Term or Bolus Exposure Limits

The question of the proper assignment of a short term or bolus exposure limits is not easy and has not received a lot of attention.  Most of us know about 8 hour and 15 minute exposure limits for occupational settings.  Some of us have even worked with instantaneous or Ceiling limits for the few chemicals for which they are assigned.  But the general question of a short term (a few seconds to say 60 minutes) exposure limits has not been heavily explored.   The question came up in last week’s blog (and offered paper) in which a spill generated a quick peak airborne and potential breathing zone concentration profile that begged for a characterization of an exposure limit relative to this peak exposure.   We deferred to the ACGIH Excursion rule.

Enter Andrew Anthony “Tony” Havics, CIH, PE (www.ph2LLC.com) who has thought about this quite a bit over the years.   Tony responded to last week's blog with the following thoughtful and, I think, very useful technical perspective which I have excerpted below:   

“The question of what concentration to use as a limit should probably be a post in itself. The ACGIH TLV concept of an excursion limit has changed over the years. The WEEL Committee currently uses the same concept as the TLV Committee. It is supported by statistics and empirical data in that if one considers a maximum 15 minute period average given an 8-hour OEL, Leidel has shown that the ratio of the 95th percentile to the mean for a GSD of 2.5 is 2.97. Since a reasonably controlled (by IH standards) workplace SEG would have a GSD of less than or = 3, the concept of 3xOEL is not a bad application. 

But one should consider whether there is a way of creating a short-term limit for a situation that has more empirical tox or biological basis. One way is to consider the endpoint. Is it an irritant with acute response? If so then the acceptable limit should not increase, e.g., OEL30min = (1)*OEL480min. Is a carcinogen? If so, then one has likely assumed a linear model of dose-response that should be constant, meaning that the acceptable limit follows Haber’s rule (actually Fluery’s rule) in that C*t = K, so OEL30min =(16)*OEL480min. Others would be somewhere in between. ten Berge proposed that (C^n)*t = K [n = 1 for carcinogens by my description]. I would suggest considering (C^n)*(t^a) = K [n = 1 & a=1 for carcinogens and n = 1, a = 0 for irritants]. Using ten Berge’s equation, one can evaluate a set of chemicals to see what the range of the variable n is typically. I evaluated 256 studies for the endpoint of death on 15 chemicals a while ago (Havics, unpublished, 2005) and found, like ten Berge, that almost all were from n= 1 to n = 3. Assuming that the model applies to other endpoints, one can estimate an equivalent OEL at 30 minutes given an OEL of say 100 ppm for 480 minutes. For n = 1, 1.5, 2, 2.5, and 3, OEL30min = 1600, 635, 400, 303, and 252 ppm, respectively; for n = 1, 1.5, 2, 2.5, and 3, OEL15min = 3200, 1008, 693, 470, and 317 ppm, respectively. So, from a conservative standpoint, the use of 3xOEL, or 300 ppm in the cases you describe seems very reasonable…” 

As a supplement, Tony sent me the following curve which graphically describes the above relationships from 480 minutes down to 7.5 minutes.  

I think the above general recommendation of n =1 or carcinogens and n = 0 for irritants is reasonably conservative given a relative lack of toxicological or biological data.  Of course, as Tony points out, this recommendation may not adequately address the needs of stakeholders assessing chemicals that are not carcinogens but could have fast-acting acute effects that might overwhelm the body's defense by bolus exposures. Here, my sense is that we (as usual) would need more data.

Using the above curves for setting exposure limits, a critical line (not on this curve but described by Tony above) is for n = 0 which would be flat at 100 ppm all the way across the curve to zero time.  This flat limit, should be considered as the exposure limit line one could use for any chemical that might cause a severe acute toxicological response where better data do not exist.   It is a conservative (i.e., overestimating) approach for any reasonably well documented 8 hour exposure limit and any chemical in which the real n is not equal to 0.

Clearly, chemicals that cause local tissue irritation, contact site toxicity and serious systemic toxicity could be lethal via acute inhalation of high concentrations even for short periods of time.   For these chemicals, those setting exposure limits really need to understand the characteristics (i.e., value of n) of the dose-response curve as a function of concentration and time of exposure. Indeed, n = 1 would not be conservative enough in many cases (see above).  Using n = 0 is, however, a reasonable default position if you have an 8 hr OEL but do not have the acute data. 

In line with these realities, Tony called to remind me that we should always be aware of the inherent uncertainty that resides in both our estimates of exposure and, in the case above, in our exposure limits.   In recent years we seem to be doing a better job of describing the uncertainty around exposure (see recent blog on MCS) but have essentially neglected this aspect on the toxicological side of risk.

The fact that these limits and all exposure limit values are not bright lines notwithstanding, we have historically assigned and used them as deterministic values.  I believe that just having the awareness that they are not bright-lines is an important first step.   Ultimately we need to quantitatively describe and control this uncertainty to assure the integrity of our assessments.

I am indebted to Tony for furthering this discussion and I would love to hear from other readers of this blog as to how we might handle do a better job of providing exposure limits for bolus exposure.