There is a range of exposure times one might consider for acute community exposure guidelines. ERPG values focus on a 60-minute exposure period. This decision was based on the availability of relevant toxicity information and an estimate of the exposure period that would be most useable for emergency response and planning scenarios. The ERPG committee has and will establish ERPGs for shorter periods upon request when there are significant reasons and data that justify a shorter exposure period.
For exposure durations of less than 60 minutes, use of ERPG values would be conservative and should be safe.

NOTE: Extrapolation to higher guidance levels for shorter exposure periods should not be attempted by use of the Haber relationship (expressing the product of exposure concentration and exposure duration as a constant), or modifications thereof, without specific validating data. The Haber relationship, with or without some of the proposed modifications, does not hold true over more than small differences in exposure time.(3) The ten Berge(4) extrapolation method can be used if there are sufficient data and expertise available to guide its implementation.
ERPGs are determined on a case-by-case basis. Different chemicals will have different dose-response curves, can cause a wide variety of health effects, and will have different amounts and quality of toxicological data. There is no fixed formula for determining ERPG values, and no fixed relationship or ratio between the three ERPG values (ERPG-1, ERPG-2, ERPG-3) for any given chemical.
Development of ERPG values is through a weight-of-evidence approach resulting from review and deliberation on the available body of data. ERPG values are developed by comprehensive review of published and proprietary original source toxicological literature. In developing an ERPG for a chemical, it is important to emphasize the use of acute or short-term exposure data. When evaluating adverse health effects, both immediate and delayed health effects are considered. When adverse reproductive, developmental, or carcinogenic effects might be caused by a single exposure, the data are considered in the derivation of the ERPG.
Human exposure data are emphasized to the extent data are available. Unfortunately, human exposure data for acute chemical exposures are often anecdotal with unknown, estimated, or reconstructed levels of exposure. As a result, animal test data often form the basis for ERPG values.
The most pertinent information is derived from acute inhalation toxicity studies in animals that include analytical determination of exposure concentrations, clinical observations, and histopathology. The focus is on the highest experimental levels not showing the effects described by the definitions of the ERPG levels. In experimental data, the methods of concentration determination (i.e., nominal vs. analytical) are important considerations.
Data from repeated inhalation exposure studies, with clinical and pathologic examinations, are also considered. When inhalation toxicity data are unavailable or limited, data from studies involving other routes of exposure (such as ingestion or dermal) will be considered. More weight is given to rigorously conducted studies.
Finally, if mechanistic or dose-response data are available, these are applied as appropriate. In every data set, data considerations such as the method to determine concentration (e.g., nominal, analytical, modeling) and duration of exposure can be important.
When appropriate and when the data are sufficient, uncertainty factors are used when determining values for ERPG-1, ERPG-2, and ERPG-3. For all three ERPG levels, a default uncertainty factor of 10 is applied for interspecies extrapolation when it is appropriate, based on the weight of evidence found in data. Lesser factors may be applied if justified by sufficient data. Conversely, additional factors may be applied when the data are insufficient or when there are unusually sensitive members of the general population (e.g., a specific metabolic defect that makes some individuals unusually susceptible to the toxicity of the substance under consideration).
To evaluate possible carcinogenic effects resulting from a single exposure to a carcinogen, the procedure described by the National Research Council (NRC), is used. If the data show the potential for carcinogenicity from long-term exposures to a chemical, the q1* calculation is performed. The q1* calculation consolidates risk estimates derived from low-dose extrapolation of animal bioassay or epidemiologic data into a single 1-hour exposure time frame and assumes a 1 in 10,000 risk of cancer. Where appropriate, physiologically-based pharmacokinetic approaches may be used to derive the risk estimates.
Users of this document are cautioned that there is no known or accepted threshold level at which persons sensitized to a chemical can be exposed without potentially experiencing an adverse skin- or respiratory-related effect. Therefore, the ERPGs may not offer protection for those who are sensitized to these kinds of chemicals.
Uncertainty factors that may be used in the derivation of ERPGs are described below. These uncertainty factors are not absolute, but rather help to develop initial discussion points from which ERPGs may be derived.
The starting point for lethality considerations almost always begins with acute mammalian toxicity data. In examining such data, uncertainty factors must always be based on quality of the data, and therefore will be highly specific for the chemical under evaluation. Adequate evaluation of animal lethality data as they apply to humans includes a weight-of-evidence approach that requires considerable experience with such data and professional judgment. This evaluation is also dependent on the depth and quality of the data set.
Over the years of setting ERPGs, an empirical relationship was noted that the predicted threshold of human lethality (ERPG-3) consistently ranged near the value of 1/30 of the 1-hr LC(5), particularly for irritants. As more data were generated, it has been noted that the highest non-lethal level in animal studies ranged near the value of 1/3 of the 1-hr LC50. The application of a 10-fold interspecies uncertainty factor results in a total factor of 30 applied to the 1-hour LC50.
For less irritating chemicals, an uncertainty factor less than 30 might be applied to the 1-hr LC50. In other cases, the uncertainty factor might be greater than 30 because of a paucity of data or the poor quality of the data available for review.
Uncertainty factors of the type used for ERPG-3 have not been developed for ERPG-2 or ERPG-1 values. ERPG-3 values are based solely on lethality and ERPG-2 and ERPG-1 values can be based on several endpoints that may not be linked mechanistically.
An odor detection indicator (yes or no) is shown with ERPG-1 values in the ERPG Table to indicate those chemicals that are likely to be detected by odor near their ERPG-1 value. This information is primarily intended for those emergency response agencies that incorporate into their planning the possibility that members of the public may call them when they detect an unusual chemical odor. This indicates only that a chemical will likely be detected by odor near its ERPG-1 value.
While a detectable odor may be indicated in the table for a specific ERPG-1 value, it should be understood that an ERPG-1 value can also be based on an airborne chemical concentration for a one-hour period below which only mild, transient adverse health effects are anticipated and that such health effects may or may not be associated with an objectionable or detectable odor. Furthermore, detection of an odor does not imply that a material is toxic at that level. Toxicity and odor detection have independent criteria. A substance may be toxic well below its odor threshold or exposures well above the odor threshold may be required to induce a toxic response.
Although the ERPG-1 definition incorporates the perception of a clearly defined objectionable odor, the property of an odor being objectionable is subjective, varies from one individual to another, and therefore is not often published. Data on odor threshold or detection levels are more commonly published. In the absence of information on an objectionable odor level, the Committee may use more conservative odor threshold or detection levels instead or may use these more conservative odor detection levels as a point of departure for estimating an objectionable level.
Lower Explosive Limit (LEL) warnings are also shown in the ERPG Table. These warnings serve to alert emergency managers and responders that an additional physical hazard may be present in addition to the toxicity hazards. The LEL warnings are presented where the ERPG values exceed 10%, 50%, or 100% of the LEL. In addition, an LEL column shows the LEL value in parts per million (ppm) when any of the ERPG values (also in ppm) exceeds any of the three LEL warning levels. The lack of an entry in the LEL column for any particular chemical should not be interpreted that the chemical is not flammable, but only that the ERPG values for that chemical do not exceed 10% of the LEL.
The AIHA Guideline Foundation expresses gratitude to all of the members of the Emergency Response Planning (ERP) Committee for their contributions to this publication.