Acute Toxicity Database

The database for the derivation of an acute reference concentration (RfC) includes a single exposure inhalation study with the rat (Driscoll and Hurley 1993), a repeated exposure study with the dog (Newton 1994b), and two well-conducted developmental toxicity studies in different species, the rabbit (Breslin et al. 1990b) and the rat (Sikov et al. 1981). In addition, there is a supporting two-generation reproductive study in the rat and pharmacokinetic studies following inhalation exposure. Therefore, the subcommittee considers the database for the derivation of an acute RfC to be good.

The subcommittee believes that DPR's use of a study with repeated exposures (Newton 1994b) as the critical study on which to base an acute RfC for children is conservative and ensures safety. The NOAEL from a study that uses a single exposure rather than repeat exposures is sufficient to derive an acute RfC provided that there is quantitative dose-response information, the study is conducted with the most sensitive species, and there is a sufficient database of supplemental toxicological information. However, because the rodent study with a single exposure to methyl bromide (Driscoll and Hurley 1993) resulted in a NOAEL that was three times higher than the NOAEL derived from the critical study with the dog (Newton 1994b), the rodent study would have resulted in a less conservative RfC.

With respect to the developmental studies by Breslin et al. (1990a,b), the subcommittee also considered it appropriate for determining an acute NOAEL for the assessment of the risks of acute occupational and residential exposure to methyl bromide. This is plausible given that a single gestational exposure is theoretically sufficient to produce an adverse developmental effect (EPA 1991), particularly blockage of gallbladder development (gestation day 11.5 to 12.5), which occurs over the course of approximately 24 hr in the rabbit. Furthermore, there are large numbers of women of childbearing age in the workforce. Finally, the maternal toxicity that occurred in the Breslin et al. studies (1990 a,b) should not negate the observed developmental effects because gallbladder development occurred 5 to 6 days before the dams displayed toxicity and because only a minority of the dams displayed toxicity. The lowest-observed-adverse-effect level (LOAEL) (80 parts per million (ppm)) used by DPR was based on the dose at which gallbladder agenesis, fetal weight declines, and fused sternebrae were noted. The NOAEL was 40 ppm, resulting in an RfC of 210 ppb.

Subchronic Toxicity Database

In addition to the above-cited developmental endpoints, there were also neurotoxic endpoints selected as critical effects for both the acute and the subchronic time periods. These endpoints were both from a single dog study (Newton, 1994b) in which the dogs at the lowest doses showed signs of depressed activity and the dogs at the higher doses and longer exposure periods showed severe signs of neurotoxicity. Because neurotoxic signs are a prominent feature of human methyl bromide intoxication, this neurotoxicity study in the dog appears to be reasonably selected as the critical study. The acute endpoint was for a NOAEL of 103 ppm, with human equivalent NOAELs of 45 ppm and 25 ppm for adults and children, respectively. Because these were higher than the human equivalent NOAEL calculated from the rabbit developmental study, an RfC was not calculated from this study.

The database for the subchronic studies appears to be quite extensive; there were numerous studies that DPR had an opportunity to evaluate to select a critical study for subchronic toxicity. DPR's selection appears to be appropriate in that they selected a study performed for regulatory purposes that was carefully designed and conducted according to the contract laboratory's standard operating procedures. The one drawback about this study (Newton 1994b) is that it was conducted to establish dose levels for a proposed chronic study (which was subsequently not performed) and not originally planned as a formal subchronic study. As a result, there were decisions made midway through the study by the authors to change the study design with respect to duration or dose levels. There were only four dogs of each sex per treatment group, which is a very small number of replications. The critical observation was made outside the standard operating procedures by a trained veterinarian on two female dogs only, which leaves the observation somewhat equivocal. However, to be conservative, the subcommittee agrees that this still appears to be the most suitable critical study out of a total of 26 studies presented as possibilities by DPR. On the other hand, the subcommittee notes that if these observations on the two dogs at 5 ppm are not considered real manifestations of methyl-bromide-mediated neurotoxicity, selection of another study (e.g., Sikov et al. 1981) as the critical one would raise the NOAEL by approximately an order of magnitude, that is, to 20 ppm (see Table 2–1).

The subchronic estimated LOAEL of 5 ppm is equivocal because of the lack of a dose-response curve at the lower dose levels, the observation of depressed activity in two of eight dogs outside the standard protocol procedures, and the low number of replications. However, the seriousness of the neurotoxicity observed in humans and the potential long-term nature of the neurological effects makes this equivocal observation reasonable as a conservative endpoint.

Chronic Toxicity Database

The existing database identified by DPR for derivation of an RfC for chronic toxicity includes two well-conducted chronic studies with different species, supported by subchronic studies in several species, a two-generation reproduction study, other data on developmental and reproductive effects, and pharmacokinetic studies employing inhalation exposure. The subcommittee considers the database available for derivation of a chronic RfC for methyl bromide good and neither of the key studies had major inadequacies.

The chronic LOAEL (3 ppm) used by DPR was based on the lowest dose that caused changes in the olfactory epithelium in rats exposed to methyl bromide for 29 months (Reuzel et al. 1987, 1991). No effects were observed in the tracheobronchial or pulmonary regions of the respiratory tract and no other exposure-related effects were noted at this concentration. The LOAEL was 30 ppm for all other more adverse effects. The NOAEL and LOAEL for respiratory effects, and also for all other effects, in the NTP study (1992) with mice were 33 ppm and 100 ppm, respectively. The critical endpoint selected from the Reuzel et al. study (1987) is appropriate as more pronounced nasal lesions have been observed at higher concentrations in shorter-term studies (Eustis et al. 1988; Hurtt et al. 1988); however, the critical endpoint in this case is observed at increased incidences only in aged rats, making it an equivocal endpoint. The Hurtt et al. (1988) study indicated that the lesions observed after exposure to methyl bromide at 200 ppm, 6 hr/day for 105 days, were largely reversible. Another consideration is the endpoint of reduced growth of neonatal rats. The NOAEL and LOAEL for reduced growth of neonatal rats in the two-generation reproduction study by American Biogenics Corporation (1986) were 3 and 30 ppm, respectively. Therefore, the Reuzel et al. study (1987) has the lowest LOAEL of the studies considered appropriate for derivation of the RfC. The subcommittee agrees with DPR's choice of this endpoint, with the notation that at the 3-ppm concentration the effects are mild and increased incidences (but not necessarily severity) are observed only in aged rats.

Developmental Neurotoxicity

Methyl bromide is clearly a neurotoxicant in human adults; neurotoxic signs are prominent following high-level human exposures and one study suggests that mild neurotoxic effects might also occur at low levels (Anger et al. 1986). Methyl bromide also is a developmental toxicant as indicated by laboratory animal studies. Therefore, there is reason to suspect that methyl bromide could be a developmental neurotoxicant, which suggests that data from a developmental neurotoxicity test would be informative to the risk assessment. However, the subcommittee finds that the developmental neurotoxicity test, as it is currently described in the U.S. Environmental Protection Agency (EPA) guidelines (EPA 1991), might be inadequate to identify and characterize specific developmental neurotoxicants. Therefore, the utility of data from such a test for DPR's regulatory needs is unclear, and the subcommittee finds that the risk characterization conducted on the currently available database by DPR is probably sufficient for identifying appropriate NOAELs.

Appropriateness of Acute-Exposure Definition

DPR's use of an acute (24-hr) exposure period is more reasonable for the residential exposure scenario than for an acute occupational (8-hr) exposure scenario. Some individuals, such as infants, young children, or elderly persons, might indeed spend most of a given 24-hr period inside the residence. However, it is unlikely that a worker will be exposed for 24 hr. For the occupational acute-exposure scenario, a shorter duration approximating the length of a work shift (8 hr) would have been more appropriate. This is particularly true for the exposure assessments and the margin of exposure analyses. For example, from the acute neurotoxicity study in dogs (Newton et al. 1994a,b, summarized in Table 3 of DPR 1999, p. 42), it can be seen that a 24-hr exposure to 50 ppm would not be the toxicological equivalent of a 6-hr exposure to 200 ppm. In dogs, a 24-hr exposure to 50 ppm is well tolerated, whereas a 6-hr exposure to 200 ppm would likely cause acute neurological signs. This becomes even more problematic for very short exposures. The 24-hr time-weighted average of a 1-hr exposure to 1,200 ppm is also 50 ppm, but based on the dog neurotoxicity study and the LC₅₀ data presented in Table 1 of the DPR report (p. 35), this is likely to be a lethal exposure for at least some of the animals. The subcommittee believes that humans would not respond differently from laboratory animals in this regard.

As a practical matter, because DPR normalized both the methyl bromide concentrations from the exposure assessment studies and the methyl bromide concentrations from the toxicity studies to 24 hr, the end effect might be that the two cancel each other out for the occupational exposure scenario, and the result might lead to a more conservative risk assessment for the residential exposure scenario. This is because the studies that DPR used to determine the critical NOAELs for acute toxicity all used exposure durations of 6 to 8 hr. DPR then normalized the NOAEL concentrations to 24 hr using concentration/exposure duration relationships. Thus, the actual exposure durations used in the studies were good approximations for an acute occupational exposure, and the two normalizations essentially canceled one another out when the margins of exposure (MOEs) were calculated. On the other hand, as already mentioned, a 6-hr exposure to 200 ppm is likely to be more acutely toxic than a 24-hr exposure to 50 ppm. Thus, the 24-hr normalized NOAEL might be lower than a NOAEL derived from actual 24-hr exposures would be. Because the residential exposure measurements should be made over 24 hr, and therefore would not have to be normalized, the MOEs for the residential exposure scenarios would be more conservative than the occupational MOEs. As discussed in Chapter 3 of this report, the fact that few actual exposure measurements were made for the residential exposure scenarios is a separate problem with the risk assessment.

Subchronic Exposure

DPR defines two categories of subchronic exposure: short-term and seasonal. The subcommittee agrees with DPR that it is appropriate to have a subchronic exposure category to describe worker exposures in preplant soil fumigation and commodity fumigation, and that a subchronic category might also be appropriate for residents of fumigated houses or residents who live near fumigation facilities. As outlined in Section IV of the DPR report (Table 19, p. 105) residents might have short-term exposures by virtue of moving back into a fumigated house. Residents may also have seasonal exposures as a result of living near fumigation facilities. The subcommittee also believes that it is plausible that residents living near fumigated fields might be exposed to methyl bromide for periods lasting longer than 24 hr, and therefore, that Section IV should include an exposure assessment for short-term and seasonal exposures to residents near fumigated fields.

The subcommittee does not believe that it is appropriate to assume, based on the short half-life of unmetabolized methyl bromide, that the effects of methyl bromide are completely reversed a few days after cessation of exposure. The subcommittee bases this statement on the fact that toxicology studies suggest that longer exposures are associated with lower NOAELs than shorter studies, indicating that some processes involved in methyl bromide toxicity are not likely to be quickly (within a few days) or completely reversible. Therefore, the subcommittee concurs with DPR that the seasonal exposure category is an appropriate one for workers who have repeated exposures to methyl bromide, separated by periods up to several days, over the course of a season.

As stated in Table 15 (DPR 1999, p. 92) and the description of the exposure durations in the DPR report (DPR 1999, p. 10), the durations for the short-term and seasonal scenarios appear to be 1 week and 6 weeks, respectively. These are the treatment durations at which effects were observed in the two critical subchronic studies identified by DPR (Sikov et al. 1981; Newton et al. 1994b). However, elsewhere in the document, DPR states that the seasonal exposure duration is “greater than one month” (DPR 1999, p. 90), and still elsewhere as 90 days (DPR 1999, Section IV.B, p. 93). The subcommittee believes that the appropriate duration is 90 days because the seasonal uses noted above are likely to last longer than 1 month or 6 weeks. The distinction between the duration of the critical studies and the actual durations of exposure for the scenarios should be clarified in Tables 15 to 20. The subcommittee concurs with the point made in the DPR report (p. 90) that the seasonal NOAEL might have been lower if the dogs in the critical study had been exposed for longer than 6 weeks (Newton et al. 1994b). However, as already discussed in this report, the subchronic RfC derived from that study is a fairly conservative one, and therefore, probably protective even for longer exposure durations.

Chronic Exposure

Chronic exposure generally refers to a 70-year (lifetime) continuous exposure to the chemical of concern. There does not appear to be chronic nonoccupational exposures for any populations associated with field agricultural applications, because the application of methyl bromide is on a seasonal basis, not year round. However, the subcommittee believes that chronic nonoccupational exposures could be possible for residents near commodity-fumigation facilities or transport facilities, where fumigation of commodity storage warehouses or shipping containers might occur on a year-round basis. Fumigation workers also might have chronic exposures.

The subcommittee believes that DPR's normalization of the 6 hr/day, 5 days/wk, exposure of the lifetime study for rats to a 24-hr/day lifetime exposure for humans is appropriate, but notes that it adds another layer of conservativeness to the derived value. The RfC should be applied to lifetime exposures. The subcommittee disagrees with DPR's definition of chronic exposure for humans as “a year or more” (DPR 1999, p. 4), because this definition does not agree with the accepted EPA definition (EPA 1989) of chronic exposure as a period between 7 years (approximately 10% of a human lifetime) and a lifetime. Subchronic exposures are defined by EPA as ranging from several months to several years.

Categories with More than Two Data Points

DPR has summarized acute exposures for worker categories when there are more than two data points as the range of the data directly observed (after adjustments for such things as application rates) plus a simple arithmetic mean and arithmetic standard deviation. However, DPR did not appear to calculate a consistent percentile from the observed data. In the current analysis, DPR has implicitly treated with greater conservatism cases in which there are more data points (a higher percentile is used) than cases in which there are fewer data points. Moreover, basing calculations on the highest of N values introduces statistical instabilities into the analysis. Finally, DPR has not provided a rationale for their choice of arithmetic means and arithmetic standard deviations, rather than more typical lognormal statistics. Analyses of data for some groups (e.g., copilots, applicators, tarp removers) by the subcommittee (see Appendix C) indicates that in general lognormal distributions would be more appropriate than would normal distributions, as is usual for exposure distributions. (For further discussion of the use of lognormal distributions for describing the variability in exposures, see Cullen and Frey (1999); Hattis and Burmaster (1994); Thompson (1999).)

Categories with One or Two Data Points

In most cases, DPR has calculated a mean (if there were two points) and listed the higher of two points as the high value. However, in some other cases, a “95^th percentile” value is calculated by assuming an arithmetic standard deviation equal to the mean of the one observed data point. DPR does not explain why this is done for some single-point exposure categories and not others. For cases in which there are only one or two data points, the subcommittee encourages DPR to either gather additional data or consolidate related exposure categories on an a priori basis (i.e., not based on the measured levels but based on similarity of the processes generating the exposures) to assemble greater numbers of data points for basic statistical treatment within groups.

Derivation of Reference Concentrations

DPR has developed inhalation RfCs for acute, subchronic, and chronic exposures. When derived from NOAELs, the RfCs reflect 100-fold uncertainties, with a 10-fold uncertainty for species differences and a 10-fold uncertainty for variations among humans. When NOAELs were estimated from LOAELs, a 10-fold uncertainty was used. DPR states that it is their policy to use a default 10-fold uncertainty factor to estimate a NOAEL from a LOAEL. The subcommittee in general agrees with this application of a default uncertainty factor of 10, particularly where the endpoint is an adverse effect such as neurotoxicity or developmental toxicity. In the case of the chronic RfC, the LOAEL was so mild as to be close to a NOAEL. In that case, the subcommittee suggests that a three-fold uncertainty factor be considered. However, the subcommittee realizes that DPR is constrained to have a chronic RfC no higher that the EPA's chronic RfC of 1.3 ppb. Therefore, DPR's chronic RfC for adults of 2 ppb is reasonable. The subcommittee also agrees that the interspecies and intraspecies uncertainty factors of 10 (for each) applied to the acute and subchronic RfCs are appropriate.

The subcommittee notes that DPR used an older method than EPA's current method for the derivation of its RfCs (EPA 1994). DPR derived separate values for adults and children; this is not possible with the current EPA method. It is interesting to note that although different methodologies were used, the RfCs derived by DPR for adults and children, 2 ppb and 1 ppb, respectively, are similar to EPA's value of 1.3 ppb.

Sensitive Subpopulations

The Food Quality Protection Act of 1996 mandated that the EPA use an additional 10-fold safety factor for infants and children, unless it could be determined from available data that a different factor would be safe. The subcommittee considered the methyl bromide database in light of the three criteria used by EPA to determine the safety factor. The first criterion concerns the completeness and reliability of the toxicology database. As discussed at length in Chapter 2 and earlier in this chapter, the subcommittee finds that the toxicology database for methyl bromide was good overall. The second criterion concerns the completeness and reliability of the exposure database. As discussed in Chapter 3, the exposure database, though flawed, is quite extensive for occupational exposures. In contrast, for residential exposures, the category into which most exposures to children and infants would fall, the database is inadequate. Limited data are available only for residential exposures during fumigation of the residence, not for residents living next to fumigated fields or fumigation facilities. The final criterion concerns the potential for prenatal and postnatal toxicity. The two-generation rat reproduction study (American Biogenics Corporation 1986) and the rabbit and rat developmental toxicity studies (Breslin et al. 1990a,b; Sikov et al. 1981) indicate that methyl bromide is not a potent teratogen, but that it can cause developmental toxicity. The teratogenic effect, gallbladder agenesis, is considered to be a minor malformation and this effect was seen only at doses that caused maternal toxicity. The subcommittee expects that a potent teratogen would cause multiple malformations at doses that do not cause maternal toxicity. As noted in the DPR report, there is some evidence for increased sensitivity of the developing organism to adverse effects of methyl bromide compared with the mothers in rats (American Biogenics Corporation 1986), but not in rabbits (Breslin et al. 1990a,b). Although good otherwise, the reproductive and developmental database lacks a developmental neurotoxicity study. According to DPR (DPR 1999, p. 126), EPA has added an additional uncertainty factor of 3 in the absence of such a study in its recent time-limited tolerances for pesticides.

Given that the NOAELs used for the various exposure scenarios are already quite conservative, the subcommittee felt that an additional safety factor for infants and children was not necessary.

Multiple Exposures

Although DPR acknowledges that workers might receive multiple exposures from methyl bromide, there is only a limited discussion on the potential exposure of residents who live in areas where multiple fields might be fumigated simultaneously or within a short period of time. Because the majority of methyl bromide is used in field applications, residents near treated fields are subject to frequent exposures during the fumigation season. The subcommittee notes that it would be unrealistic to assume that most residents in agricultural areas live near only one treated field. Therefore, the buffer zones established by DPR to be protective of residents adjacent to one field might not be sufficient should the residents be near multiple treated fields. Although these exposures were commented upon in Chapter 3, the subcommittee reiterates that this is a significant data gap in the exposure assessment.

In addition, the subcommittee has concerns regarding repeated exposures of workers, such as soil or structural fumigators, because soil fumigators might have repeated exposures on consecutive days for several months or structural fumigators might be engaged in multiple fumigations on a single day (Anger et al. 1986). The potential that such repeated exposures might occur raises concerns in light of results from Anger et al. (1986) that suggest that relatively low exposure levels (<2 to 3 ppm) of methyl bromide from fumigation might produce slight neurotoxic effects in workers. Additional data on the neurotoxic effects of methyl bromide in exposed workers are needed.

The proposed regulations provide for a 36-hr waiting period between the application of methyl bromide to a field near a school and when school is in session. In practical terms, this means that fumigation of fields near schools are limited to Friday evenings and Saturday. However, DPR makes no provision for school activities that might occur during weekends at the school, particularly outdoor activities such as sports. Such exposures should not be ignored, because children might have greater susceptibility to effects from methyl bromide exposures, and because data suggest that slight neurotoxic effects might occur at low concentrations (Anger et al. 1986).