134a-exposed group. However, the significance of these results is uncertain because the study did not involve lifetime exposure.

The NRC also evaluated an inhalation study by Hext and Parr-Dobrzanski (1993), which was subsequently published by Collins et al. (1995). In that study, rats (85 of each sex per group) were exposed (whole body) to HFC-134a at concentrations of 0, 2,500, 10,000, or 50,000 ppm for 6 hr per day, 5 days per week for 104 weeks. The only exposure-related effect of toxicological importance was an increased incidence of Leydig-cell hyperplasia and Leydig-cell adenoma in male rats in the 50,000-ppm group. The tumors were benign and not life threatening. The survival rate was similar in all groups. In 1996, the NRC concluded that Leydig-cell tumors were not applicable to humans and thus were not considered an adverse effect.

Two additional studies have been published. Alexander et al. (1995) exposed (nose-only) mice (60 of each sex per group) to HFC-134a at concentrations of 2,500, 15,000, or 75,000 ppm for 1 hr per day for at least 104 weeks. There were two air-only (control) exposure groups. Clinical observations, behavioral observations (Modified Irwin Screen), body weights, hematology, and microscopic tissue pathology were monitored. There were no exposure-related effects in any of the measurements.

In another study, rats (60 of each sex per group) were exposed (nose only) to HFC-134 at concentrations of 2,500, 10,000, or 50,000 ppm for 1 hr per day for at least 108 weeks (Alexander et al. 1995). HFC-134a vapor was delivered using a metered-dose inhaler. An air-only (control) exposure group consisted of 120 rats of each sex. Clinical observations, behavioral observations (Modified Irwin Screen), body weights, hematology, and microscopic tissue pathology were monitored. A statistically significant increase in subacute and chronic laryngitis occurred in female rats only in the 50,000-ppm group. The severity of laryngitis was slight, and the observation was considered to be of no toxicological significance. There were no exposure-related effects in any of the other measurements.

In 1996, the NRC reviewed the available toxicity data on HFC-134a and proposed a 1-hr EEGL of 4,000 ppm, a 24-EEGL of 1,000 ppm, and a 90-day CEGL of 900 ppm. The Navy chose to set lower values for the 1-hr EEGL and 90-day CEGL because of the lack of experience with HFCs, but it did adopt the NRC's proposed 24-EEGL.

Since the 1996 review, additional data on HFC-134a have become avail-

able. An updated summary of the noncancer toxicity studies on HFC-134a are presented in Table 4-4. The subcommittee used the new studies to reevaluate the exposure guidance levels proposed by the NRC in 1996 and those currently used by the Navy (see below). Because the submariner population is all male, young, and healthier than the general public, the committee did not use an uncertainty factor for intraspecies variability in its calculations.

One of the new studies was an ascending-concentration safety study in humans. The subjects were exposed to HFC-134a at concentrations up to 8,000 ppm for 1 hr with no adverse effects (Emmen and Hoogendijk 1999). The subcommittee believes that this study should be used to determine the 1-hr EEGL rather than the cardiac-sensitization study in dogs (Hardy et al. 1991) used by the NRC (1996) in its earlier evaluation of HFC-134a. Although the human subjects were not challenged with epinephrine as in the dog study, the subcommittee notes that the human NOAEL of 8,000 ppm is five-fold lower than the NOAEL of 40,000 ppm for dogs. Thus, the subcommittee believes that a 1-hr EEGL of 8,000 ppm can be justified. That value is four-fold greater than that currently used by the Navy.

The subcommittee considered a 13-week toxicity study in rats (Hext 1989; Collins et al. 1995) to be the most appropriate for deriving the 24-hr EEGL for HFC-134a. In this study, the highest concentration of 50,000 ppm was the NOAEL. Because the available data on HFC-134a were inadequate to determine the magnitude of difference between rats and humans, the NOAEL was divided by an uncertainty factor of 10 to account for interspecies variability, which yielded a 24-hr EEGL of 5,000 ppm. This exposure level is higher than the 1,000-ppm guidance level used by the Navy and proposed by the NRC in 1996. The reason for the difference is that in 1996 the NRC was determining exposure levels for use aboard Navy ships with female crew members and, therefore, based the NOAEL of 10,000 ppm on a devel-

opmental study in which fetal toxicity was observed (Hodge et al. 1979a). However, fetal toxicity is not as a relevant an end point for setting an exposure level for use on submarines, which have no female crew members.