Acute inhalation toxicopathology of lithium combustion aerosols in rats

FUNDAMENTALANDAPPLlEDTOXICOLOGY

7,58-67(1986)

Acute inhalation Toxicopathology

of Lithium Combustion

Aerosols in Rats’

A. H. REBAR,* B. J. GREENSPAN,~ AND M. D. ALLEN Inhalation Toxicology Research Institute, Lovelace Biomedical and Environmental Research Institute, P.O. Box 5890, Albuquerque, New Mexico 87185

Acute Inhalation Toxicopathology of Lithium Combustion Aerosols in Rats. REBAR, A. H.. GREENSPAN, B. J., AND ALLEN, M. D. (1986). Fundam. Appl. Toxicol. 7,58-67. Male and female F344/Lov rats were exposed to aerosols produced by burning lithium metal under conditions designed to stimulate a fire in the containment building of a fusion reactor. Lithium combustion aerosols were generated by sweeping lithium vapor into air atmospheres with controlled CO? and HZ0 concentrations. Chemical analyses of the aerosols produced indicated a dependence of the chemical form of carbon dioxide concentrations and relative humidity. Under conditions of low CO2 concentration and low relative humidity (<25%), the aerosol was predominantly lithium monoxide with some lithium hydroxide and about 12% lithium carbonate. Under conditions of high relative humidity (>75%), the aerosol was primarily lithium hydroxide with about 23% lithium carbonate. Although these two aerosols might be expected to have different acute toxicities based on their differing alkalinities, the 14day LC50 values (with 95% confidence limits) determined after 4-hr exposures were 940 (730-1200) mg/m3 for the lithium oxide and hydroxide mixture and 960 (830-1200) mg/m3 for the lithium hydroxide aerosols. Histopathologic lesions were observed in the nasal turbinates, larynx, and occasionally in the lungs with both aerosols. The most prominent lesions were necrotizing laryngitis and ulcerative rhinitis. Pulmonary lesions represented a secondary extension of the upper respiratory tract lesions rather than a primary manifestation of lithium toxicity. The similarities in the LC50 valus and also in the histopathologic lesions observed suggest that any Liz0 in the aerosol reacted rapidly with water vapor in the respiratory tract to form LiOH prior to deposition.

Current interest in the toxicity of inhaled lithium combustion aerosols was stimulated by proposalsto utilize lithium metal in fusion reactors and spacenuclear power systems.Lithium will be usedto breed tritium to fuel fusion reactors. Also, becauseit is easily pumped and has a high specific heat, lithium may be used asa heat transfer fluid (Young and Gore, 1976) in fusion and space nuclear reactors. In fact,

the primary heat transfer loops of a typical fusion power plant may contain up to lo6 kg of molten lithium metal (Gore and Murphy, 1976).All pipestransporting lithium metal will be enclosed inside a containment vesselwith an inert atmosphere. Although unlikely, there is a potential for an accidental releaseof large quantities of molten lithium metal into an air atmosphere. This would result in a fire and possibly a large releaseof lithium combustion aerosols. Jeppson (1982) has experimentally demonstrated that the concentration of lithium aerosol produced by a pool fire within a containment building could reach as high as 16,000 mg/m3. If molten lithium metal accidently comes into contact with air, it reacts exothermically with all of the major constit-

’ Research supported by the Office of Health and Environmental Research of the U.S. Department of Energy under Contract DE-AC04-76EV0 10 13. ’ Present address: Department of Veterinary Microbiology, Pathology and Public Health, School of Veterinary Medicine, Purdue University, West Lafayette, Ind. 47907. 3 Present address: Biology and Chemistry Department, Battelle Pacific Northwest Laboratory. P.O. Box 999. Richland, Wash. 99352. 0272-0590186

$3.00

58

TOXICOPATHOLOGY

uents. This results in a rapid temperature increase and the production of a dense cloud of aerosol. Lithium initially reacts with oxygen to form lithium monoxide (Li*O). Then, if water vapor is present in the atmosphere, lithium monoxide will react to form lithium hydroxide (LiOH). Lithium hydroxide is chemically more stable but less caustic than lithium monoxide. If CO2 is present in the atmosphere, lithium hydroxide will react to form lithium carbonate (Li2C03). Lithium carbonate is chemically more stable but less caustic than lithium hydroxide. Nothing has been published about the reaction kinetics of lithium combustion aerosols. In a lithium fire, the composition of the aerosol would depend on the Hz0 and CO;! content of the atmosphere. If a fire occurs in an enclosed containment vessel, it is possible for the COz concentration to become depleted so that the aerosol has a higher percentage of the more caustic forms of lithium (i.e., Liz0 and LiOH). In a fusion power plant there is a potential for a concurrent rupture of a steam line inside the containment vessel. If steam were injected into the containment during a lithium fire, the aerosol would be composed primarily of lithium hydroxide monohydrate (Jeppson, 1982). Although the toxicity of orally administered lithium salts is relatively low (Schou, 1976), combustion aerosols produced by a lithium fire and inhaled might be expected to be extremely caustic to the respiratory tract. The pathologic alterations seen following acute exposures to aerosols of related sodium combustion products have been well documented (Busch ef al., 1983; Zwicker et al., 1979). Prevalent lesions were acute laryngitis, laryngeal edema, and multifocal hemorrhage in the trachea and lungs (Busch et al., 1983; Zwicker et al., 1979). In an earlier paper, we reported the results of an acute inhalation toxicity study using an aerosol that was approximately 80% lithium carbonate, the remainder being lithium monoxide and lithium hydroxide (Greenspan et al., in press). This aerosol was intended to

OF

LITHIUM

59

AEROSOLS

simulate the situation which would result if lithium aerosol were released outside of the containment structure to the ambient atmosphere where it would be exposed to an unlimited supply of COz. Under these conditions, the aerosol would be expected to be predominantly lithium carbonate. The purpose of this research was to investigate the acute inhalation toxicity of lithium aerosols that could be formed under distinct sets of conditions: (1) a major lithium fire inside a containment vessel in which the humidity is low (~25%) and the CO* concentration has been depleted by reactions with lithium and (2) a major lithium fire inside a containment vessel in which the CO* is depleted and the humidity is high (>75%). The goal of this study was to determine the 14-day LC50 values and pathologic alterations for rats exposed to lithium aerosols composed of either predominantly lithium hydroxide or predominantly lithium monoxide, and to compare the results with those from our previous inhalation studies using aerosols that were predominantly lithium carbonate. MATERIALS

AND

METHODS

Animal selection and housing. Four groups of 8 male and 8 female laboratory-raised F344/Lov rats. aged 9 to 12 weeks, were used in the two series of lithium exposures. A single control group consisting of the same number of animals was shared between the two lithium exposure series. The animals were housed two per cage in filter-topped polycarbonate plastic cages (48 X 24 X 20 cm high) with hardwood chip bedding. Rooms were maintained at 6872”F, with a relative humidit of 20-50% and a 12-hr light: dark cycle with light starting at 6:00 AM Food (Lab Blox, Allied Mills, Chicago, Ill.) and water from bottles with sipper tubes were provided ad libitum. The animals were randomly assigned into exposure groups by body weight. Any animal appearing in poor health or having an initial body weight beyond two standard deviations of the group mean for each sex was not used in the study. Exposure regimen. The two series of exposures in the present study consisted predominantly of LiOH and a mixture of Liz0 and LiOH, respectively. Both exposure aerosols contained limited (up to 25%) LizC03. The predominantly LiOH aerosol was generated under conditions of high relative humidity (>75%) and the mixture of Liz0 and LiOH was generated under conditions of low relative humidity (~25%). The animals were exposed to the lithium

60

REBAR,

GREENSPAN.

combustion aerosols once for 4 hr in a I-m3 Rochestertype chamber. The target concentrations for both series of exposures were 500, 750, 1000, and 1500 mg/m3 and were selected to bracket the expected LC50 values. The control group of animals was exposed in the chamber to air only. Each series of exposures was performed within a 2-week period. The present exposures were designed to generate data for comparison with those generated in our previous L&CO3 exposures which have been detailed in a previous report (Greenspan et al., in press). The lithium carbonate exposure results have been summarized along with data from the present studies in Tables l-3. The control animals for the Li*CO, exposure were done in conjunction with the Li2C03 exposure and consisted of 16 animals exposed to room air in the exposure chamber. The control data listed in Table 3 are combined data for both the L&CO3 exposures and the present exposures. .4erosolgerzerurion. The lithium aerosol generation system has been described in detail previously (Allen et al.. in press). A stainless-steel melting chamber filled with lithium metal was heated inductively to temperature up to 1300°C. depending on the desired aerosol concentration. Ultrapure argon flowed through the melting chamber at 7 liters/min and carried lithium vapor into an air atmosphere having controlled CO1 and H20 concentrations. Lithium metal vapor reacted with air to form an intense, white flame that produced dense aerosol concentrations. These aerosols flowed through the animal exposure chamber at 200 liters/min. The mass concentration of aerosol in the chamber was monitored by drawing filter samples (0.8pm Acropore filters, Gelman Instrument Co.. Ann Arbor, Mich.) every 15 min throughout the exposure. The filters were weighed with an electrobalance (Model 27, Cahn Instruments. Cerritos. Calif.). An optical aerosol monitor (Model P-5A, Environmental Monitoring Systems, Knoxville, Term.) provided real-time monitoring of the aerosol concentration in the chamber. Adjustment ofthe melting chamber temperature provided excellent control over the chamber concentration. Cascade impactor samples (Mercer ef al., 1970) were taken three times during each exposure to determine the aerodynamic particle size distribution of the aerosol. The quantity of lithium collected on each stage was assayed with a flame photometer (Model C-5 1. Coleman Instrument Co., Oakbrook, Ill.). Continuous monitoring of temperature (Type E chromel-constantan thermocouple), relative humidity (Model HS-ICHDT-2A digital humidity monitor. Thunder Scientific, Albuquerque, N. Mex.), oxygen concentration (Model 245R O2 monitor, MSA. Pittsburgh. Pa.), and CO* concentration (Miran Model 101 Infrared Analyzer. Wilks Scientific. Norwalk, Conn.) was performed in the exposure chamber. During each 4-hr exposure, three aerosol samples were collected in deionized water with a glass impinger to determine the chemical composition of the aerosol by dilute

AND

ALLEN

acid titration. Ten-minute samples were collected at a flow rate of 6 liters/min. After aerosol was collected with the impinger, it was transferred to a clean 100-ml beaker and titrated with 0.05 N HCl. The pH of the solution was continuously monitored with a pH meter connected to a loin. strip chart recorder. The mass of Liz0 or LiOH and the mass of LIzCO~ in the solution were computed from the volume of HCI required to bring the solution to inflection points corresponding to pH values of 8.5 and 4.0. Using this technique, the difference between Liz0 and LiOH cannot be distinguished quantitatively. This method only measures the mass of LizCOl and the mass of Liz0 or LiOH that was collected in the impinger. However. since the reaction of the Liz0 to LiOH in the presence of water vapor is rapid, it can be assumed that the aerosol is primarily LiOH . Hz0 for lithium combustion in humid atmospheres. Clinical observations. Body weights were measured on Days 0 (preexposure), 1. 5. 8, 14, 21, and 28. Clinical examinations for signs of toxicity (respiratory difficulty, encrustation of blood or mucus around the eyes, nose, or mouth. righting reflexes, etc.) were performed daily. SacriJce and necropsy. At I4 days after exposure. a portion of the surviving animals were selected randomly for sacrifice by ip injection of a lethal dose of T-6 I euthanasia solution (American Hoechst Co., Sommerville, N.J.). The remaining animals were observed for an additional I4 days and then sacrificed. A complete necropsy was performed on all animals. Lung. liver. kidney, and thymus weights were recorded. The thymus, liver, kidneys, and respiratory tract, from nares to lungs, were removed and fixed in 10% neutral buffered Formalin for histopathology. Tissue sections stained with hematoxylin and eosin were prepared for light microscopy. Data analysis. The 14-day LC50 values were calculated for the male and female rats separately and combined using probit analysis (Finney, 1964). Because the exposures were not concurrent, and because animals died at various times after exposure. only qualitative comparisons could be made for body weight and organ weight data.

RESULTS Aerosol Characteristics The aerosol particles showed the typical branched-chain structure of vaporizationcondensation aerosols of metal oxides. The average mass median aerodynamic diameter (MMAD) for the LiOH exposure series was 0.70 k 0.14 pm (SD, n = 8) with an average geometric standard deviation of 1.55 + 0.22 (SD). For the Liz0 plus LiOH exposure series, the average MMAD was 0.61 k 0.07 pm (SD,

TOXICOPATHOLOGY

OF

n = 12) with an average geometric standard deviation of 1.46 + 0.10 (SD). The null hypothesis that the average MMAD values measured for these two exposure series were equal could not be rejected at a significance level of a = 0.05. This result is an indication that aerodynamic size was independent of chemical form in these experiments. Table 1 summarizes the results of titration measurements of the chemical composition of lithium aerosols generated under various conditions of relative humidity and CO2 concentration. The values listed in Table 1 are the arithmetic mean and standard deviations of the results from three series of animal exposures. Case 1 lists the results from a series of exposures reported earlier (Greenspan et al., in press), in which the predominant aerosol produced was lithium carbonate. Case 2 gives the results from a series of exposures in which lithium metal vapor was mixed with an atmosphere having a low CO* concentration and a high relative humidity; these conditions produced an aerosol that was predominantly iithium hydroxide. Case 3 reports the results from a series of exposures in which lithium vapor reacted with an atmosphere having a low CO* concentration and a low relative humidity; the resulting aerosol was a mixture of lithium monoxide and lithium hydroxide.

LITHIUM

61

AEROSOLS

Acute Toxicity Within the first 10 min of exposure, numerous animals exhibited excessive salivation and nasal mucus secretion. Coughing and choking were observed throughout the exposure. Daily observations of the animals showed respiratory difficulty, including audible mouth breathing, encrustation of blood and mucus around the eyes and nose, slowing of the righting reflex, and general lethargy; these observations correlated well with increased exposure concentrations. Several animals had swollen, inflammed external genitalia. Rapid and persistent weight loss was characteristic of animals in the higher exposure concentration groups. Animals exposed to only 500 mg/m3 exhibited transient and reversible weight losses. Table 2 shows the mortality data from exposures designed to determine the acute LC50 of the aerosols in cases 2 and 3 of Table 1. These aerosols were generated as predominantly lithium hydroxide (case 2) and as a mixture of predominantly lithium monoxide with some lithium hydroxide and some lithium carbonate (case 3). The combined male and female 14-day LC50 values (with 95% confidence limits) for the two exposure series were 960 (830-l 200) mg/m3 for case 2 and 940 (730- 1200) mg/m3 for case 3. In both sets

TABLE

1

EFFECT OF RELATIVE HUMIDITY AND CO2 CONCENTRATION ON THE CHEMICAL COMFQSITIONOF LITHIUM AEROSOLS

Case 1 2 3

Relative humidity (%I

Concentration (ppm)

36 I 6 78 -t- 2 25 +-4

NOW Values represent means and standard centrations for each case.

of CO2

Percentage by weight of Liz0 and LiOH

Percentage by weight of L&CO3

21.4 + 2.2 77.1 -r- 9.4 88.7 * 5.1

78.6 t 4.4 23.0 f 9.2 12.0 k 3.6

2300 t 290 170+ 50 40* 5 deviations

of three

measurements

made at each of four exposure

con-

62

REBAR,

GREENSPAN,

AND

TABLE

ALLEN

2

MORTALITY DATEFOR MALEAND FEMALE RATSEXPOSEDTOLITHKJM Aerosol concentration (w/d

Composition LizCOs-Case

LiOH-Case

Liz0

1

2

and LiOH-Case

’ No surviving

3

AEROSOLS

Days postexposure Sex

0

1-3

4-7

8-14

Combined mortality through Day

620

M F

0 0

O/16

1400

M F

2 0

2116

2300

M F

4 4

14116

570

M F

0 0

1 0

0 0

5116

840

M F

0 0

2 0

0 0

2116

1200

M F

1 0

6 0

0 1

9116

1500

M F

5 0

2 3

1 4

--a

500

M F

0 0

0 0

0 1

0 0

l/16

750

M F

0 0

0 0

0

2 2

5116

1000

M F

0 0

0 0

3 0

1 2

6116

1500

M F

6 2

1 3

3

14

15/16

0

-

16116

animals.

of exposures, the male rats died earlier than the female rats exposed concurrently. No clear dose-response relation could be discerned for absolute organ weight or organ weight-to-body weight ratios in the exposed animals. Pathology

The incidence and distribution of histologic alterations are summarized in Table 3. Lesion incidence data for our previous studies with animals exposed to Li2C03 aerosols also have been included. Alterations were seen in the larynges, turbinates, and lungs of exposed animals. In the predominantly LiOH (case 2) and

L&CO3 regimens, males had a greater incidence of lesions; this pattern was less evident in the animals exposed primarily to L&O (case 3). There was a clear dose-response pattern for both males and females in all exposure groups. Laryngeal lesions were the most consistent alterations with all of the lithium aerosols tested. The principal change was necrotic laryngitis (Fig. 1). The lesion was segmental, affecting primarly the epiglottis and anterior portions of the larynx with the more distal portions often entirely spared. In affected regions there was mucosal erosion and/or ulceration, submucosal inflammation with extension into laryngeal musculature, and oc-

TOXICOPATHOLOGY

OF

LITHIUM

TABLE INCIDENCE

AND

DISTRIBUTION

OF LESIONS

Exposure level (w/m *)

Aerosol Control LizC03-Case

0 1

620 1400 2300

Total LiOH-Case

M

2

570 840 1200 1500

Liz0 and LiOHCase 3

No. animals with laryngitis/total

M

F

T

M

F

T

O/16

O/32

O/16

O/l6

O/32

0116

O/16

O/32

O/16 3116 12116

018

O/16 2116 9116

O/8

018 l/8

O/16 4116 12/14

018

O/16 3116 9116

318

318 3124

218 W3 218 O/8 418 218 4/s 218

315

414

7129

7128

018 018 O/8 618 618 9124

2116 2116 6116 6116

4132

O/8 118 118 118 318 218

No. animals with thymic lymphoid depletion

T

O/8 v3

No. animals with alveolitis/total

F

0116

018

CONCENTRATIONS

M

O/32

3/s

TO VARYING AEROSOLS

T

O/I6

12132 500 750 1000 I500

F

EXPOSED

COMBUSTION

O/16

9124

Total

Total

No. animals with rhinitis/total”

3

IN RATS

OF THREE LITHIUM

63

AEROSOLS

218 W-3 318 018 718 218 718 618 19132

l/16 2116 5/16 719

6124

3/16 9116 13116

8132

O/8 88 l/8 O/8 318 118 215 214 6129

2116

5128

2116 l/l6 4/16 419

O/8 V8 O/8 718 218

9124

2/24

v-3 l/8 6/g

O/8 O/8

3/8

318 4/s

II/32

l/32

‘38 .m 018 018 218 O/8 115

314

3129

5128

No@. Thymes was not saved at necropsy in all animals. n M = male: F = female: T = total. ’ Toxicity of LizCO> exposures has heen summarized in detail in a previous report (Greenspan

casional thrombosis of submucosal vessels. There was bacterial colonization of ulcerated surfaces and occasional laryngeal foreign bodies (hair fragments) were present. Mucosal surfaces immediately adjacent to necrotic zones frequently exhibited squamous metaplasia. The majority of animals with laryngeal pathology also had turbinate lesions. Occasional animals had turbinate lesions with no evidence of laryngeal involvement. Turbinate lesions consisted of focal to patchy erosive to ulcerative rhinitis (Fig. 2). The nasal cavity often contained an exudate formed of aggregates of neutrophils enmeshed in mucus. Areas of squamous metaplasia were found within the nasal mucosa. Bacterial colonization and occasional nasal foreign bodies (hair fragments and keratinous debris) were observed. Pulmonary lesions were of varying severity

l/16 l/l6 9116 7116

2116 O/16 2116 419

318

516

718

8122

8124

617

418

5/7 60 317

3/s

20128

14130

218 7.13

118 616 117

617 415

114 318 414

14123

9123

10/15 6115 9115 9113

3115 317 9115 s/9

ef al.. in press).

and were present only in animals with laryngeal and/or turbinate involvement. The mildest lesions were rare and consisted of acute multifocal terminal bronchiolitis and associated alveolitis. More severe lesions consisted of patchy diffuse necrotizing pneumonia, often associated with inhaled foreign bodies (hair fragments and keratinous debris) and bacterial colonies. Thymic lesions were also present. The principal change was depletion of cortical lymphocytes with resultant organ atrophy. Multifocal hemorrhage was a common accompaniment. DISCUSSION This study was designed to investigate the acute inhalation toxicity of lithium aerosols

64

REBAR,

GREENSPAN,

AND

ALLEN

FIG. 1. Photomicrograph of the larynx from a rat exposed to an aerosol of predominantly LiOH (50X). The anterior larynx is at right, the distal larynx at left and laryngeal lumen at bottom. There is necrosis of the anterior laryngeal epithelium with superficial bacterial colonization and extension of inflammation into the underlying tissues. Epithelium adjacent to the necrosis (as left) has undergone squamous metaplasia.

that might be produced during an accident at a fusion reactor. During such an event, the chemical composition of the combustion aerosols will depend upon the concentration of CO2 and water vapor in the air. In a closed containment vessel, the aerosol will be initially composed of L&O, LiOH, and Li2COJ. As the COI, concentration is depleted, the percentage of the aerosol that is lithium carbonate will be considerably reduced. Rupture of a steam line would increase the relative humidity and cause an increase in the percentage of the aerosol that is lithium hydroxide. In a previous paper (Greenspan et al., in press) we reported on the toxicity of lithium combustion aerosols that were approximately 80% Li2C03 and the remainder either Liz0 or LiOH. The measured LC50 value (with 95% confidence limits) for those aerosols was found to be 1800 ( 1600-2 100) mg/m3. For case 2 in

the present studies, in which the aerosol was intended to be primarily lithium hydroxide, the measured LC50 value was 960 (130- 1200) mg/m3. For case 3, in which the aerosol was intended to be primarily Liz0 with some LiOH, the LC50 value was 940 (730-1200) mg/m3. A higher measured LC50 for animals exposed to aerosols that were predominantly L&CO3 is consistent with the fact that lithium carbonate is the least alkaline of the lithium combustion aerosols. Even though Liz0 is more caustic than LiOH, the measured LC50 values for aerosols produced at high humidities (case 2) were not significantly different than those produced at low humidities (case 3). This observation may be due to the rapid reaction of Liz0 to LiOH in the humid environment of the upper respiratory tract. In the present studies, as with our previous studies with aerosols comprised principally of

TOXICOPATHOLOGY

OF

LITHIUM

AEROSOLS

65

FIG. 2. Photomicrograph of a turbinate scroll from a rat exposed to an aerosol of predominantly L&CO3 (50X). The scroll is central, nasal cavity is at top and bottom. The nasal cavity is filled with exudate composed of degenerating inflammatory cells and epithelial lining cells. At top, the epithelial lining has undergone necrosis and the exudate is adherent to the underlying submucosa.

Li2C03 (Greenspan et al., in press), primary morphologic lesions were restricted to the upper respiratory tract. While the character of the pathologic alterations was similar for all three aserosols, the incidence and distribution of the lesions differed (Table 3). The lesion incidence data reflected the LC50 data; Li2C03 produced both fewer and less severe lesions than the other two exposure regimens at similar concentrations. Laryngeal lesions were more consistent than turbinate lesions (Table 3) probably because exposure to the aerosols resulted in an almost immediate conversion from a normal respiratory pattern to one of mouth breathing. The type of lesion, primarily necrotizing, supports the suggestion that combustion aerosols of alkali metals exert their toxicity by producing localized areas of high pH upon contact with the tissue, thereby causing protein coagulation.

Such a pathogenesis was first proposed by workers testing the toxicity of inhaled lithium hydride (Spiegl et al., 1956). The reason for the absence of primary lesions in the deeper portion of the larynx and in the trachea is unknown. One possible explanation is that only a small portion of the aerosol penetrated beyond the anterior larynx despite the small aerodynamic particle size. Failure to penetrate deeply may have resulted from several factors. First, the rapid collection ( 1- 10 min into exposure) of foamy fluid in the naopharynx and larynx may have acted as an impinger that removed caustic lithium particles from inhaled air. Additionally, the rapid development of coughing and attendant laryngospasm may have significantly altered air flow patterns with resultant particle deposition in the nasal cavity and anterior larynx. With reduced penetrance of the aerosol, a

66

REBAR, GREENSPAN,

concentration gradient would develop from the nasopharynx to the more distal portions of the upper respiratory tract. In the nasal cavity and anterior larynx where concentrations would have been high, the buffering capacity of body fluids and exhaled CO2 may have been inadequate to neutralize the caustic effects of deposited particles. In contrast, in more distal regions of the respiratory tract, where concentrations would have been relatively lower, neutralization may have occurred. Supportive evidence for neutralization of caustic materials in the respiratory tract is provided by inhalation studies with sulfuric acid where the marked buffering capacity of the lung has been documented (Larson et al., 1979). More severe upper respiratory tract lesions featured both the presence of foreign bodies (inhaled hair fragments) in the larynx and nasal cavity and bacterial colonization of necrotic zones. Altered respiratory patterns undoubtedly favored the inhalation of foreign bodies. The loss of normal epithelial surfaces favored colonization of the lesions by the bacteria introduced with the foreign material. Pulmonary lesions in all cases represented secondary extensions of the septic upper respiratory tract lesions. There was no evidence of a direct caustic or erosive effect on pulmonary tissues. Respiratory tract pathology induced by our lithium aerosols (including Li2C03) was similar to that induced by sodium combustion aerosols and lithium hydride aerosols (Busch et al., 1983; Spiegl et al., 1956; Zwicker et al., 1979). In all of these studies, laryngeal lesions have been prevalent. In both the lithium hydride studies and the sodium studies, occasional tracheal necrosis was present; this change was not seen in our exposures. It is of interest that for the lithium hydride studies, tracheal involvement was present only in mice and not in rats, a finding which does correlate with the absence of tracheal necrosis in our rats. These data suggest that the depth of penetrance of lithium aerosols is at least partially species dependent.

AND ALLEN

A significant difference between our studies and previous investigations is in the type of pulmonary pathology observed. Pulmonary lesions were not described in the lithium hydride exposures and occurred only rarely in the sodium combustion aerosol studies (Busch et al., 1983; Spiegl et al., 1956; Zwicker et al., 1979). Even in the sodium aerosol studies, only hemorrhage, and not the fulminant secondary septic bronchopneumonia which we described, was seen. This discrepancy cannot be explained, especially in view of the isolation of Pseudomonas spp and/or Streptococcus spp from liver and/or heart blood from all of the sodium-exposed animals which were culturally examined (Busch et al., 1983). Thymic lesions in our studies, while treatment related, probably do not represent a direct toxic effect of the lithium aerosols. The principal thymic change was depletion of cortical lymphocytes, a common nonspecific reflection of systemic stress. Similarly, the apparent gender-related difference between male and female must be interpreted cautiously, While, in general, males did have a higher lesion incidence than females, they were also larger and probably inhaled a greater total dose than did females. Total dosage was probably a more important determinant of lesion incidence than gender.

ACKNOWLEDGMENTS The authors acknowledge R. B. Simpson for his technical assistancewith generating the aerosol and S. A. Likens for her assistancewith the animals. The authors are grateful for suggestions and critical review of the manuscript by Dr. J. M. Benson, Dr. W. E. Bechtold, Dr. N. A. Gillett, Dr. F. F. Hahn, Dr. S. J. Rothenberg, Dr. A. R. Dahl, Dr. J. A. Mewhinney, Dr. R. G. Cuddihy, Dr. B. B. Boecker. and Dr. R. 0. McClellan. The generation system was originally designed and constructed at the Battelle Pacific Northwest Laboratories, Richland, Washington, and was graciously loaned to the Institute for these studies. This work was performed under U.S. Department of Energy Contract DE-AC04-76EV0 10 I3 in facilities fully accredited by the American Association for Accreditation of Laboratory Animal Care.

TOXICOPATHOLOGY

REFERENCES ALLEN, M. D., GREENSPAN, B. J., BRIANT, J. K., AND HOOVER, M. D. (in press). Generation of lithium combustion aerosols for animal inhalation studies. Health Physics. BUSCH,R. H., MCDONALD, K. E., BFUANT,J. K., MORRIS, J. E., AND GRAHAM, T. M. (1983). Pathologic effects in rodents exposed to sodium combustion products. Environ. Res. 31, 138-147. KINNEY, D. J. (1964). Probit Analysis. University Press, Cambridge. GORE, B. F., AND MURPHY, E. S. (1976). Current Fusion Power Plant Design Concepts. Battelle Pacific Northwest Laboratory Report BNWL-2013, UC-20. GREENSPAN,B. J., ALLEN, M. D., AND REBAR, A. H. (in press). Inhalation toxicity of lithium combustion aerosols in rats. J. Toxicol. Environ. Health. HAMMOND, P. B., AND BELILES, R. P. (1980). Metals. In Casarett & Doull’s Toxicology (J. Doull, C. D. Klaasen, M. 0. AMDUR, eds.), Macmilian, New York. JEPPSON,D. W., BALLIF, J. L., WEAN, W. W., ANDCHAU, B. E. (1978). Lithium Literature Review: Lithium’s Properties and Interactions. Hanford Engineering Development Laboratory Report HEDL-TME-78-15. JEPPSON,D. W. (1979). Interactions of Liquid Lithium

OF

LITHIUM

AEROSOLS

67

with Various Atmospheres, Concretes and Insulating Materials and Filtration of Lithium Aerosols. Hanford Engineering Development Laboratory Report HEDLTME-79-7,

UC-20.

JEPPSON,D. W. (1982). Scoping Studies--Behavior and Control of Lithium and Lithium Aerosols. Hanford Engineering Development Laboratory Report HEDLTME-80-79.

UC-20.

LARSON, T. V., FRANK, R., COVERT, D. S., HOLUB, D., AND MORGAN, M. (1979). Measurements of respiratory ammonia and the chemical neutralization of sulfuric acid aerosol in anesthetized dogs (Abstract). Amer. Rev. Respir. Dis. 119, 226. MERCER, T. T., TILLERY, M. I., AND NEWTON, G. J. (1970). A multi-stage low flow rate cascade impactor. J. Aerosol Sri. 1, 9-l 5. SCHOU, M. (1976). Pharmacology and toxicology of lithium. Annu. Rev. Pharmacol. Toxicol. 16,23 l-243. SPIEGL, C. J., SCOTT, J. K.. STEINHARDT, H., LEACH, L. J., AND HODGE, H. C. (I 956). Acute inhalation toxicity of lithium hydride. Arch. Ind. Health 14,468-470. YOUNG, J. R., AND GORE, B. F. (1976). Reference Commercial Fusion Power Plants. Battelle Pacific Northwest Laboratory Report BNWL-2014, UC-20. ZWICKER, G. M., ALLEN, M. D., AND STEVENS, D. L. ( 1979). Toxicity of aerosols of sodium reaction products. J. Environ. Toxicol. Pathol. 2, 1139-l 150.