No reliable data on human exposures via the inhalation route were located. All AEGL values are based on a study in which rats were exposed to concentrations of 0, 10, 30, 50, 100, or 150 parts per million (ppm) for 8 or 12 h (Kim and Carlson 1986). The only reported effect was methemoglobin formation. The relationship between aniline concentration and methemoglobin formation appeared to be linear. Furthermore, the relationship between methemoglobin formation and time, between 3 and 8 h, was also linear when the aniline concentration was held constant at 100 ppm. Methemoglobin reached an asymptote at 8 h. Based on the linear relationships between aniline concentration and methemoglobin formation and between methemoglobin formation and time at a constant aniline concentration, a linear relationship between concentration and exposure duration (C1 × t = k) was chosen for time-scaling aniline concentrations to the appropriate AEGL exposure durations. Although the key study (Kim and Carlson 1986) used an 8-h exposure, methemoglobin measurements were taken at several time points during the study, and other studies with 4-h (E.I. du Pont de Nemours 1982; Pauluhn 2002) and 10-min exposures (Kakkar et al. 1992) support the derived AEGL values. Thus, the 8-h AEGL values from the Kim and Carlson study were extrapolated back to 10 min. Following a 10-min exposure, the concentration of methemoglobin in blood is unlikely to reach steady state, as typically seen 6-8 h after the initiation of exposure.
The AEGL-1 was based on an exposure of rats to a concentration of 100 ppm for 8 h, which resulted in elevation of methemoglobin from a control value of 1.1% (range, 0.4-2.1%) to 22%. A review of the published data indicates that methemoglobin levels of 15-20% in humans result in clinical cyanosis but no hypoxic symptoms. Although inhalation data for comparison purposes are not available, oral ingestion data suggest that humans may be considerably more sensitive to methemoglobin-forming chemicals than rats. Therefore, a default uncertainty factor of 10-fold was used for interspecies extrapolation (NRC 1993). Several sources also indicate that newborns may be more sensitive to methemoglobin-forming chemicals than adults. Because of the lack of specific quantitative data on sensitive human subpopulations and the fact that there are data suggesting greater susceptibility of infants, a default uncertainty factor of 10-fold was also used for intraspecies extrapolation. It is believed that an intraspecies uncertainty factor of 10 is protective of the general population, including susceptible individuals. A default uncertainty factor of 10 for each of the interspecies and intraspecies variabilities is also supported by the small database of information and the lack of reliable human inhalation studies. The data were scaled across time using C1 × t = k because of data indicating a linear relationship between concentration and exposure duration as related to methemoglobin formation. The AEGL-1 values are supported by the data of Pauluhn (2002) in which dogs exposed to 46 ppm for 4 h had the same methemoglobin concentration (4.7%) as rats exposed to 50 ppm for 8 h (Kim and Carlson 1986; at 50 ppm, methemoglobin steady state in the blood is attained after several hours but prior to the full 8-h exposure).