2.1. Acute Lethality

Although a great many deaths have occurred from accidental, intentional, or occupational exposures to HCN, in only a few cases are specific exposure concentrations known. In a review of human fatalities (ATSDR 1997), it was stated that exposure to airborne concentrations of HCN at 180 to 270 ppm were fatal, usually within several minutes, and a concentration of 135 ppm was fatal after 30 min. The average fatal concentration for humans was estimated at 546 ppm for 10 min. The latter data point is based on the work of McNamara (1976), who considered the resistance of man to HCN to be similar to that of the goat and monkey and four times that of the mouse. Fatal levels of HCN cause a brief period of central nervous system stimulation followed by depression, convulsions, coma with abolished deep reflexes and dilated pupils, and death. Several review sources, such as Dudley et al. (1942), Hartung (1994), and ATSDR (1997), report human toxicity data that appear to be based largely on pre-1920 animal data.

2.2. Nonlethal Toxicity

Several studies of occupational exposures and one study with a human subject were located. In the occupational exposures (summarized in Table 5– 3), neurological symptoms consistent with cyanide intoxication were demonstrated, but the likelihood of concomitant exposure to other chemicals could not be ruled out. For example, cleaners and cutting oils, as well as sodium and copper cyanide, may be present in electroplating operations (ATSDR 1997). The experimental human study involved the exposure of a single subject and a dog to a high concentration for a short exposure period.

TABLE 5–3

Occupational Exposures to Hydrogen Cyanide.

Adverse health consequences on systems other than the central nervous and respiratory systems have been documented during occupational and/or accidental exposures. Generally, these effects occurred following chronic exposures, but the cardiovascular and dermal effects could occur following acute exposures. For example, cardiovascular effects (palpitations, hypotension, and chest pain) (El Ghawabi et al. 1975; Blanc et al. 1985; Peden et al. 1986), hematological effects (increased or decreased hemoglobin) (El Ghawabi et al. 1975; Kumar et al. 1992), hepatic effects (increased serum alkaline phosphatase activity but not serum bilirubin) (Kumar et al. 1992), gastrointestinal effects (nausea and vomiting) (El Ghawabi et al. 1975), endocrine effects (thyroid enlargement) (Hardy et al. 1950; El Ghawabi et al. 1975; Blanc et al. 1985), and dermal effects (burns and rashes) (Blanc et al. 1985; Singh et al. 1989) have been observed. Authors of several studies, including Hardy et al. (1950), observed that some of the symptoms of chronic cyanide exposure are a result of thiocyanate-induced goiter. These authors noted that goiter has also been reported following thiocyanate therapy for hypertension.

El Ghawabi et al. (1975) compared the symptoms of 36 workers exposed to HCN in three electroplating factories in Egypt with a referent group; employment ranged between 5 and 15 y. None of the workers in either the exposed or control groups were smokers. Cyanide exposure resulted from a plating bath that contained copper cyanide, sodium cyanide, and sodium carbonate. Concentrations of cyanide in the breathing zone of the workers ranged from 4.2 to 12.4 ppm (means in the three factories: 6, 8, and 10 ppm). Fifteen-minute air samples were collected in NaOH and analyzed colorimetrically. Symptoms reported most frequently by exposed workers compared with the referent control group were, in descending order of frequency: headache, weakness, and changes in taste and smell. Lachrimation, vomiting, abdominal colic, precordial pain, salivation, and nervous instability were less common. The authors made no attempt to correlate the incidences of these symptoms with concentrations. Although there were no clinical manifestations of hypoor hyperthyroidism, 20 of the workers had thyroid enlargement to a mild or moderate degree; this conditions was accompanied by higher ¹³¹I uptake compared with the referent controls. Exposed workers also had significantly higher blood hemoglobin, lymphocyte cell counts, cyanmethemhemoglobin, and urinary thiocyanate levels than controls. Urinary thiocyanate levels were correlated with cyanide concentration in workplace air. Two workers in the factory with a mean exposure of 10 ppm suffered psychotic episodes; recovery occurred within 36 to 48 h. Although the sample size was small, the study used well-matched controls and included a biological index of exposure (urinary thiocyanate). The NRC Subcommittee on Spacecraft Maximum Allowable Concentrations, in evaluating the El Ghawabi et al. (1975) data, concluded that “8 ppm would likely produce no more than mild CNS effects (e.g., mild headache) which would be acceptable for 1-hour exposures” of healthy adults (NRC 2000). ATSDR (1997) noted that exposure to cleaners and cutting oils may have contributed to the effects observed in this study.

Grabois (1954) surveyed HCN levels in five plants that processed apricot kernels in order to determine possible health hazards. The survey was performed by the Division of Industrial Hygiene of the New York State Department of Labor. Work area concentrations in the plants ranged from <1 to 17 ppm, and two areas in one of the plants had levels of 17.0 ppm (comminuting area) and 13.9 ppm (cooking area). The general workroom atmosphere in this plant averaged a 6.4 ppm concentration of HCN. Medical questionnaires were not given and the health status of the employees was not reported. However, recommendations were made for controlling HCN exposures “where required,” presumably where concentrations were above the then maximum recommended concentration of 10 ppm. NIOSH (1976), in interpreting the Grabois (1954) data, stated that 5 ppm was a no-effect level, and higher concentrations were only rarely present.

Chandra et al. (1980) studied the effects of HCN exposure on 23 male workers engaged in electroplating and case hardening. The workers avoided cyanogenic foods such as cabbage and almonds for 48 h prior to blood and urine sampling. In spite of the low exposure levels—0.8 mg/m³ (0.7 ppm) in the breathing zone and 0.2 mg/m³ (0.2 ppm) in the general work area—the workers complained of typical symptoms of HCN poisoning (symptoms not stated); however, no objective measures of adverse health effects were reported. Higher blood and urine cyanide and thiocyanate were measured in exposed workers compared with a control group. Higher levels of blood and urine cyanide and thiocyanate were present in smokers than in nonsmokers in both the exposed and control groups.

Blanc et al. (1985) surveyed and examined 36 former employees of a silver reclaiming facility in order to determine acute and potential residual adverse health effects resulting from occupational HCN exposure. The study was prompted by a worker fatality from acute cyanide poisoning. The workers had been exposed long-term to excessive concentrations of cyanide as the time-weighted average (TWA) taken 24 h after the plant had closed down was 15 ppm. The most frequently reported symptoms included headache, dizziness, nausea or vomiting, almond or bitter taste, eye irritation, loss of appetite, epistaxis, fatigue, and rash. The most prevalent symptoms (headache, dizziness, nausea or vomiting, and a bitter or almond taste) were consistent with acute cyanide poisoning. A concentration-response relationship corresponding to high- and low-exposure jobs was demonstrated, but exact breathing zone concentrations were unknown. Some symptoms exhibiting a dose-response trend occurring seven or more months after exposure had ceased. Mild abnormalities of vitamin B₁₂, folate, and thyroid function were detected and suggested long-term cyanide and thiocyanate involvement. The NRC (2000), in reviewing this study, pointed out that the 24-h TWA of 15 ppm was measured one day after the plant had closed down, suggesting that workers may have been exposed to cyanide at more than 15 ppm.

Hardy et al. (1950) observed increased urinary excretion of thiocyanate in a group of case-hardener workers (hot metals are dropped into baths of cyanide salts in order to harden the material). Two workers with unqualified exposures suffered persistent headaches, sweating, chest pains, dizziness, fatigue, weakness, mental confusion, disturbed motor function, nervousness, coughing, sneezing, cramping in the lower abdomen, auricular fibrillation, and thyroid enlargement. The authors indicated that ≤10 ppm should prevent cyanide toxicity in workers, and with adequate engineering controls, workers were routinely exposed at 4–6 ppm. No symptoms were surveyed or discussed for these routine exposures.

Radojicic (1973) reported fatigue, headache, weakness, tremor in the arms and legs, pain, and nausea in 28 electroplating workers and 15 foundry workers chronically exposed to cyanide. Employment duration ranged from 0 to 15 y. Area atmospheric concentrations ranged from 6 to 13 ppm in the electroplating facility (four measurements) and 5 to 8 ppm in the foundry (three measurements). In the electroplating facility, higher concentrations were measured over work vats, 10 to 13 ppm, than in the middle of the room where concentrations were 6 to 8 ppm. In both facilities, urinary thiocyanate levels of workers were higher after work than prior to work, were higher in smokers than in nonsmokers, and increased with the number of years of work. Urinary thiocyanate concentrations were higher in smokers prior to a work day than in nonsmokers following a work shift. Symptoms were more pronounced in workers with the longer exposures.

Urinary and blood cyanide and thiocyanate were measured in a group of 140 workers consisting of exposed and nonexposed smokers and nonsmokers (Maehly and Swensson 1970). The HCN-exposed group consisted of 39 nonsmokers and 55 smokers. Area measurements, sampled with Draeger tubes at each work station, ranged from 1 to 10 ppm (average, 5 ppm). Blood and urinary cyanide and thiocyanate levels varied widely among the groups, and there was no clear relationship to occupational exposure at these concentrations; blood cyanide levels did not bear a relationship to exposure via smoking, but free thiocyanate levels in the urine tended to be higher in smokers than in nonsmokers. No worker symptoms were reported in this study.

Leeser et al. (1990) reported a cross-sectional study of the health of cyanide-salt production workers. Sixty-three cyanide production workers employed for 1 to 40 y were compared with 100 referent workers from a diphenyl oxide plant. Workers were examined before and after a block of six 8-h shifts. All workers had full medical examinations, routine clinical chemistry tests, and blood samples taken for measurement of blood cyanide and carboxyhemoglobin. In addition, circulating levels of vitamin B₁₂ and thyroxine (T4) were measured. Atmospheric cyanide was monitored with static monitors, Draeger pump tests, and personal monitoring. For the personal monitoring, air was drawn through bubblers which contained sodium hydroxide. Cyanide collected in the sodium hydroxide solution was measured using an anion-selective ion electrode. All results (a total of 34 samples) were between 0.01 and 3.6 mg/m³ (0.01 and 3.3 ppm). Geometric mean values for eight job categories ranged between 0.03 and 1.05 mg/m³ (0.03 and 0.96 ppm). Values for only one job category (eight personal samples) averaged 0.96 ppm. Results of routine Draeger pump tests (area samples) were between 1 and 3 ppm (none were above 10 ppm). In addition, during the fall of the year, production problems in part of the plant caused the HCN level to increase to 6 ppm from the usual 1–3 ppm (measurement method not stated). This increased exposure was reflected in an increase in mean blood cyanide level in the workers following a block of six 8-h shifts, and there was an increase of 5.83 μmol during the 6 ppm exposure compared with a decrease of 0.46 μmol across the shift block in the spring. Static monitors on all floors, set to trigger alarms at 10 ppm, did not sound during the study. Blood cyanide levels in exposed workers, though low, were generally higher than in control workers, and the highest levels were measured in cyanide-exposed nonsmokers compared with the nonsmoking control group (cyanide-exposed nonsmokers, 3.32 μmol; controls, 1.14 μmol; p<0.001). For ex-smokers, the difference was smaller (cyanide exposed, 2.16 μmol; controls, 1.46 μmol), and for current smokers, the blood cyanide level was actually higher in the control group (2.94 μmol for cyanide workers who smoked; 3.14 μmol for controls who smoked). The percentage of workers reporting symptoms such as shortness of breath and lack of energy was higher in cyanide workers than in the diphenyl oxide plant workers. These differences were partially explained by the greater number of cyanide workers who were shift workers. Slightly higher hemoglobin values and lymphocyte counts in the cyanide workers were not dose-related. Results of clinical and physical examinations and evaluation of medical histories failed to reveal any exposure-related health problems.

A 20-year-old man employed in a photographic darkroom suffered attacks of numbness, weakness, vertigo, some nausea, rapid pulse, and flushing of the face after 1 h of work (Parmenter 1926). Two other workers were unaffected. Following improved ventilation in the room, cyanide was measured in several areas of the workroom, including over a sink into which ferrous sulphate and potassium cyanide were routinely disposed. Concentrations of cyanide at that time (with the improved ventilation) ranged from 25 to 75 ppm.

During inspection of a tank containing a thin layer of hydrazodiisobutyronitrile (HZDN), a worker collapsed after 3 min, was fitted with a breathing apparatus after another 3 min, and removed from the tank after 13 min, resulting in a 6-min exposure (Bonsall 1984). At that time the worker was unconscious with imperceptible breathing and dilated pupils. He was covered with chemical residue. The tank had previously been washed with water; HZDN decomposes with water to give HCN and acetone. No HCN was measured prior to entry into the tank, but immediately after the incident, levels of HCN of about 500 mg/m³ (450 ppm) were measured. One hour after the exposure, the comatose individual was administered sodium thiosulfate, and following subsequent complications and treatment, he was discharged after 2 weeks (wk). No sequelae were apparent.

Barcroft (1931) described the controlled exposure of a 45-year-old, 70-kg man and a 12-kg dog to a concentration of HCN at 500–625 ppm in an airtight chamber. The human volunteer attempted to maintain the same level of activity as the dog. The dog became unsteady at 50 seconds (s), unconscious at 75 s, and convulsive at 90 s. One second later, the man walked out of the exposure chamber with no apparent effect. At 5 min after initiation of exposure, the man experienced a momentary feeling of nausea, and at 10 min from the start, his ability to concentrate in “close conversation” was altered. The dog at first appeared to be dead but recovered without adverse signs by the next day. Barcroft (1931) cites two other studies in which fumigation workers were exposed to a concentration of HCN at 250 ppm for 2 min or 350 ppm for 1.5 min without dizziness.

2.3. Developmental and Reproductive Effects

No data concerning developmental or reproductive effects of HCN in humans were identified in the available literature.

2.4. Genotoxicity

No data concerning the genotoxicity of HCN in humans were identified in the available literature.

2.5. Carcinogenicity

No data concerning the carcinogenicity of HCN in humans were identified in the available literature.

2.6. Summary

A great many human fatalities associated with acute HCN exposure have occurred, but exposure concentrations are for the most part unknown. Acute exposures that failed to result in mortality were either to high concentrations for very short exposure durations (approximately 500 or 450 ppm for approximately 1.5 min or 6 min, respectively [Barcroft 1931; Bonsall 1984]) or to exposure concentrations and times that were estimated (>25 ppm for about 1 h [Parmenter 1926]). Monitoring studies indicate that workers were routinely exposed at ≤10 ppm (Hardy et al. 1950; Grabois 1954; Maehly and Swensson 1970). Occupational HCN exposures at 1–10 ppm were acceptable at the time of these surveys as 10 ppm was the maximum acceptable concentration for workers. More effective exhaust ventilation was implemented “where required,” presumably where exposures were greater than 10 ppm, as in the Grabois (1954) study. The low exposures in the Leeser et al. (1990) study did not result in adverse health effects. Concentrations greater than 8–10 ppm may cause discomfort, and with long-term exposures, more serious symptoms can develop (El Ghawabi et al. 1975). The most common complaints in the monitoring study by El Ghawabi et al. (1975) were headache, weakness, and changes in taste and smell. Specific exposure levels for specific symptoms were not provided nor were concurrent exposures to other chemicals noted. Chronic exposure to low concentrations of HCN has been associated with hypothyroidism (development of goiter) (Hardy et al. 1950), and some symptoms associated with chronic exposures may be attributed to thyroid effects. It should be noted that in the study of Radojicic (1973) symptoms in workers increased with the number of years of work, and 20 of 36 workers in the study of El Ghawabi et al. (1975) had thyroid enlargement. No information on developmental and reproductive effects, genotoxicity, or carcinogenicity in humans was located.

3.1. Acute Lethality

Acute inhalation lethality data for the rat, mouse, and rabbit for exposure times of 10 s to 12 h were located. A single inhalation study with the dog did not give an exposure duration. The data are summarized in Table 5–4. Data from studies with nonlethal concentrations are summarized in Table 5–5. Barcroft (1931) reported LC₅₀ values and times to death for eight species of animals, the times to death at a constant concentration. Due to experimental design constraints, the LC₅₀ values are not reported here, but relevant data are discussed in the section on relative species sensitivity (Section 4.4.1).

TABLE 5–4

Summary of Acute Lethal Inhalation Data in Laboratory Animals.

TABLE 5–5

Summary of Nonlethal Inhalation Data in Laboratory Animals.

3.2. Nonlethal Toxicity

Toxicity studies resulting in nonlethal effects are reported in Table 5–5. Acute exposure data were available for the monkey, rat, and mouse with exposure durations ranging from 5 min to 24 h. Limited data were available for the dog.

3.3. Developmental and Reproductive Effects

No information regarding developmental and reproductive effects of HCN in animals via the inhalation route was located in the available literature. The teratogenic potential of inorganic cyanide was studied by infusing sodium cyanide to pregnant golden hamsters between gestation days 6 and 9 (Doherty et al. 1982). Anomalies were induced only at tested doses of 0.126 to 0.1295 mmol/kg/h because preliminary tests had shown that a dose of 0.125 mmol/kg/h did not produce anomalies, and a dose of 0.133 mmol/kg/h produced 100% resorptions. Maternal signs of toxicity were observed after 36 to 48 h, at which time the doses administered by infusion were 30 to 40 times the subcutaneous LD₅₀. This range of doses produced high incidences of congenital malformations and resorptions. The most common anomalies were neural tube defects including encephalocoele and exencephaly. Fetal crown-rump length was significantly reduced in the offspring of treated dams. Maternal toxicity did not correlate with the incidence of anomalies in the offspring. Simultaneous subcutaneous infusion of thiosulfate protected against the teratogenic effects of cyanide. Signs of cyanide poisoning appear if detoxification occurs at a slower rate than absorption (90% of an acute lethal dose of cyanide can be detoxified in an hour when given to guinea pigs by slow infusion). Because signs of maternal toxicity did not appear for 36 to 48 h, the authors suggested that the rate at which sulfur in the form of thiosulfate, cystine, or cysteine became available for cyanide detoxification was the critical step. In addition to sodium cyanide, aliphatic nitriles and cyanogenic glycosides have been demonstrated to be teratogenic to golden hamsters by the oral and inhalation routes (Willhite 1981, 1982; Willhite and Smith 1981; Willhite et al. 1981; Frakes et al. 1985, 1986a,b). The teratogenic activities were attributed to the cyanide released through metabolism of the parent compounds; in each case, developmental toxicity was observed only at doses also inducing signs of maternal cyanide intoxication.

3.4. Genotoxicity

No information regarding the genotoxicity of HCN in animals was located in the available literature. Studies that addressed genotoxicity from other forms of cyanide were reviewed in ATSDR (1997). In those studies, cyanide in the form of potassium cyanide tested negative in Salmonella typhimurium strains TA1535, TA1537, TA1538, TA98, TA10, TA97, and TA102; one study gave positive results with strain TA100. Sodium cyanide gave negative results in several strains of S. typhimurium. Potassium cyanide also tested negative in the DNA repair test in Escherichia coli and in an in vivo testicular DNA synthesis inhibition test with the mouse.

3.5. Chronic Toxicity and Carcinogenicity

No information regarding the carcinogenicity of HCN in animals via the inhalation route was located in the available literature. In a 2-y feeding study, ten male and ten female rats were administered food fumigated with HCN at each of two concentrations (Howard and Hanzal 1955). The average daily concentrations were 73 and 183 mg CN/kg diet. Based on food consumption, body weight, and concentrations at the beginning and end of each feed preparation period, estimated doses were 4.3 and 10.8 mg CN/kg body weight per day. There were no treatment-related effects on body weight and no clinical signs or histopathologic lesions attributable to cyanide ingestion. In a review of feeding studies by the U.S. Environmental Protection Agency (EPA) (1993), 10.8 mg/kg/d (11.2 mg/kg/d as HCN), in the study by Howard and Hanzal (1955), was identified as the highest NOAEL.

3.6. Summary

Lethality data were available for the rat, mouse, and rabbit for exposure periods of 10 s (rat) to 12 h (mouse). Five-minute LC₅₀ values ranged from 323 ppm (mouse) to 503 ppm (rat). Thirty-minute LC₅₀ values ranged from 166 ppm for the mouse to an average of 177 ppm for the rat. The average 1-h LC₅₀ value for the rat was 134 ppm. The LC₅₀ values tend to be similar for the mouse and rat, and the mouse was slightly more sensitive in accordance with its slightly smaller body size and higher relative respiratory rate. Sublethal effects were characterized by incapacitation (or loss of consciousness) and changes in respiratory or cardiac parameters. Exposures causing little to no effect were: monkey, 60 ppm for 30 min—slight changes in EEGs; rat, 200 ppm for 12.5 min—changes in cardiac-released blood enzymes; rat, 55 ppm for 30 min—changes in pulmonary parameters; and mouse, 63 ppm for 30 min—50% decrease in respiratory rate. No information on developmental and reproductive effects, genotoxicity, or carcinogenicity by the inhalation route was located in the available literature. Genotoxicity studies with cyanide salts were generally negative, and no cancers were induced in rats in a 2-y feeding study with HCN.

4.1. Metabolism and Disposition

HCN is miscible with water and is taken up by the moist respiratory passages. Retention levels of HCN in the nose and lung of human subjects were measured by Landahl and Herrmann (1950) while the subjects inhaled 0.5 to 20 ppm. HCN was delivered to the nose via a mask; the sample was drawn through the nose and out of the mouth while the subject held his breath. Using this procedure, the percentage retained in the nasal passages ranged from 13% to 23%. The percentage retained by the lung when inhaling through the mouth (no mask) ranged from 39% to 77%. The average exposure time was 1 min.

HCN in the blood is almost completely contained in the red blood cells where it is bound to methemoglobin. Immediately after infusion of sodium nitroprusside into patients, 98.4% of the blood cyanide was found in the red blood cells (Vesey et al. 1976). At normal physiological levels of body methemoglobin (0.25% to 1% of the hemoglobin), a human adult can bind about 10 mg of HCN (Schulz 1984).

HCN is detoxified to thiocyanate (SCN⁻) by the mitochondrial enzyme rhodanese; rhodanese catalyzes the transfer of sulfur from thiosulfate to cyanide to yield thiocyanate, which is relatively nontoxic (Smith 1996). The rate of detoxification of HCN in humans is about 1 μg/kg/min (Schulz 1984) or 4.2 mg/h, which, the author states, is considerably slower than in small rodents. This information resulted from reports of the therapeutic use of sodium nitroprusside to control hypertension. Rhodanese is present in the liver and skeletal muscle of mammalian species as well as in the nasal epithelium. The mitochondria of the nasal and olfactory mucosa of the rat contain nearly seven times as much rhodanese as the liver (Dahl 1989). The enzyme rhodanese is present to a large excess in the human body relative to its substrates (Schulz 1984). This enzyme demonstrates zero-order kinetics, and the limiting factor in the detoxification of HCN is thiosulphate. However, other sulfur-containing substrates, such as cystine and cysteine, can also serve as sulfur donors. Other enzymes, such as 3-mercapto-pyruvate sulfur transferase, can convert cyanide to thiocyanate (ATSDR 1997; NRC 2000). Thiocyanate is eliminated in the urine.

Venous blood levels of cyanide reached a steady state (mean value, 200 μg/100 mL) within 10 min of exposure of cynomolgus monkeys at 100–156 ppm (Purser et al. 1984). The blood level stayed constant during the remainder of the 30-min exposure, during which time the animals lost consciousness; the blood level remained the same for 1 h after exposure, even though the monkeys recovered consciousness within 10 min. The mean concentration of whole blood cyanide in rabbits that died following inhalation exposure was 170 μg/100 mL; the mean plasma concentration was 48 μg/100 mL (Ballantyne 1983).

Plasma levels of cyanide in unexposed, healthy adults average 0 to 10.7 μg/100 mL (mean, 4.8 μg/100 mL) (Feldstein and Klendshoj 1954). Following mild exposures to cyanide, plasma levels return to this normal range within 4 to 8 h after cessation of exposure; the half-life for the conversion of cyanide to thiocyanate from a nonlethal dose in humans was between 20 min and 1 h.

Although Feldstein and Klendshoj (1954) reported plasma levels of cyanide, most data available are for whole blood. Average whole blood values for cyanide are as follows: nonsmokers, 1.6 μg/100 mL; smokers, 4.1 μg/100 mL; and nitroprusside therapy, 5 to 50 μg/100 mL (Tietz 1986). These data can be compared with the whole blood values measured in several studies, including the study of Aitken et al. (1977) in which patients were infused with nitroprusside solutions to induce hypotension during intracranial surgery (see Box 1–1). In the Chandra et al. (1980) study, blood cyanide levels of up to 220 μg/100 mL appear excessively high in light of the low measured exposures. Snodgrass (1996) states that blood cyanide greater than 20 μg/100 mL may be associated with acute signs of cyanide poisoning, and deaths occur after blood cyanide reaches 100 μg/100 mL. As noted by Aitken (1977), metabolic acidosis occurred in patients at blood cyanide levels of ≥90 μg/100 mL.

BOX 1–1

Whole Blood Levels of Cyanide in Monitoring and Nitroprusside Therapy Studies. Leeser et al. (1990) Chandra et al. (1980)

It should be noted that HCN can be absorbed through the skin. For this reason, ACGIH (1996) and NIOSH (1997) guidelines carry a skin notation. Drinker (1931) cites the case of three men protected with gas masks in an atmosphere of 2% (20,000 ppm) HCN. After 8 or 10 min the men felt symptoms of marked dizziness, weakness, and throbbing pulse. They left the chamber just before collapse. For several hours after the exposure they experienced weakness, high pulse rate, and headache. They were incapacitated for several days, followed by complete recovery. Based on exposure to several cyanide salts, the dermal LD₅₀ in rabbits was calculated to be 6.7 mg CN⁻/kg (Ballantyne 1983).

4.2. Mechanism of Toxicity

HCN is a systemic poison that acts on the central nervous system. HCN interrupts cellular respiration by blocking electron transfer from cytochrome oxidase to oxygen. Tissue oxygen levels rise, resulting in increased tissue oxygen tension and decreased unloading for oxyhemoglobin. As a consequence, oxidative metabolism may slow to a point where it cannot meet metabolic demands. This is particularly critical in the brainstem nuclei where lack of an energy source results in central respiratory arrest and death. Cyanide can inhibit many other enzymes, particularly those that contain iron or copper, but cytochrome oxidase appears to be the most sensitive enzyme. Cyanide also stimulates the chemoreceptors of the carotid and aortic bodies to produce a brief period of hyperpnea. Cardiac irregularities may occur, but death is due to respiratory arrest (Hartung 1994; Smith 1996). Brain lesions have been associated with exposure of animals to high concentrations of HCN (ATSDR 1997).

Wexler et al. (1947) studied the effect of intravenously administered sodium cyanide on the electrocardiogram of 16 soldiers. A dose of 0.15 to 0.2 mg/kg (HCN at 0.06–0.11 mg/kg) was chosen based on the known inability of 0.11 mg/kg to stimulate respiration during medical tests (a dose of 0.11 mg of sodium cyanide per kilogram of body weight is used to determine arm-tocarotid blood circulation time). The electrocardiograms of 15 of the 16 men revealed a sinus pause (without auricular activity), which persisted for 0.88 to 4.2 s. The sinus pause immediately preceded or accompanied respiratory stimulation. The pause was followed by marked sinus irregularity, a slowing of the heart rate for a few seconds to 2 min, followed by a gradual acceleration to rates above the baseline level. Baseline heart rate and rhythm were generally restored within 3 min. There was a lesser effect on the sixteenth subject. According to AIHA (2000), this dose is equivalent to inhaling 10 ppm for 1 h.

4.3. Structure-Activity Relationships

No structure-activity relationships were applicable for establishing AEGLs for HCN. It has been observed that the signs of intoxication associated with excessive exposure to HCN and with certain aliphatic nitriles are similar. While the toxic concentrations of acrylonitrile are similar to HCN when compared on the basis of cyanide content (Dudley et al. 1942), the time course of aliphatic nitrile intoxication is different. The authors also observed that dogs are more susceptible to acrylonitrile than monkeys, but repeated exposures to acrylonitrile were more toxic to monkeys than to rats, guinea pigs, or rabbits.

4.4. Other Relevant Information

4.4.1. Species Differences

Lethal concentrations are relatively similar for various animal species and humans (Hartung 1994), with the monkey and goat being the least sensitive, according to Barcroft (1931). Barcroft (1931) reports relative species sensitivity as determined by time to death (in minutes) at a concentration of 1,000 mg/L (910 ppm): dog, 0.8; mouse, cat, and rabbit, 1.0; rat and guinea pig, 2.0; goat, 3.0; and monkey, 3.5. He reported that monkeys (two monkeys per exposure) were only beginning to show signs of unsteadiness when the dogs (two dogs per exposure) died. Also, Barcroft's study (1931) with one human subject and one dog tends to indicate that dogs are much more sensitive to the effects of HCN than humans. Barcroft notes that body size and respiration rate influence the rapidity of effect, small, rapidly respiring animals succumbing first, but he also notes that there are exceptions to the body size effect (i.e., the goat was much less sensitive than the dog). Bancroft's pre-1970 animal studies were not cited in Section 3.1 because time to death is not useful in determining exposure concentration-duration relationships but is useful for determining relative species sensitivity.

Relative to body weight, humans have a much lower respiratory rate and cardiac output than rodents. These are the two primary determinants of systemic uptake of volatile chemicals. Therefore, at similar nominal concentrations, rodents absorb substantially more cyanide than primates. From a pharmacokinetic view, lower hepatic rhodanese levels in primates will not be significant at high, acute HCN exposures. It should be noted that Barcroft's subject withstood a 1 min and 31 s exposure at approximately 500 to 625 ppm without immediate effects (Barcroft 1931), whereas mice suffer asphyxia during a 2 min exposure at 500 ppm (Matijak-Schaper and Alarie 1982). Compared with rodents, the respiratory tracts of humans and monkeys are more similar in gross anatomy, the amount and distribution of types of respiratory epithelium, and airflow patterns (Barrow 1986; Jones et al. 1996).

In the rat and mouse studies by Higgins et al. (1972) and the rat and rabbit studies by Ballantyne (1983), LC₅₀ values differed by less than a factor of two (1.5). All of the 30-min LC₅₀ values summarized in Table 5–4 range from 157 to 200 ppm (rat, mouse, and rabbit and excluding the restrained rats in the study by Levin et al. [1987]). The 1-h LC₅₀ values range from 120 ppm to 144 ppm (data for rat only). The LC₃₀ for the rat at 6 h was 68 ppm. The LC₃₀ and LC₅₀ values are presented graphically in Figure 5–1. The concentrations for the rat are means for the respective time intervals. As can be seen in Figure 5–1, the concentration-time curve is steep, particularly at the shorter time intervals.

FIGURE 5–1

Lethality values for three species of animals. All values are LC₅₀ values except the data point for the rat at 360 min, which is an LC₃₀.

Species differences are recognized in the activity of rhodanese; sheep have relatively high levels of activity and dogs have relatively low levels (Aminlari and Gilanpour 1991). Himwich and Saunders (1948) assayed tissues from several animal species for their ability to produce thiocyanate from cyanide. Activity was generally highest in liver tissue. Rats had the highest levels, dogs had the lowest levels, and rhesus monkeys and rabbits had intermediate levels. Liver and kidney rhodanese activity was two to three times higher in rats and hamsters than in rabbits and female beagles (Drawbaugh and Marrs 1987). The authors point out that in acute exposures at high concentrations, the normal low levels of rhodanese present in tissues would not allow time for substantial detoxification, and other pharmacokinetic considerations may be important in the outcome of acute poisonings.

4.4.3. Concentration-Exposure Duration Relationship

When data are lacking for desired exposure times, scaling across time may be based on the relationship between concentration and exposure duration (Cⁿ ×t=k) when a common end point is used (ten Berge et al. 1986). The end points for HCN are incapacitation and lethality. Regression analysis of the data of Sakurai (1989), using incapacitation concentrations for mice for the exposure durations of 5, 10, 20, and 30 min, results in a value for n of 1.6. Regression analysis of the incapacitation data of Purser et al. (1984) for monkeys for the time period of 8 to 19 min results in a value for n of 2.1 (Appendix A, Figure A-1). These studies were of relatively short duration.

FIGURE A-1

Regression line for incapacitation in monkeys (data of Purser et al. [1984])

Several lethality studies conducted over various exposure durations were available for calculation of concentration-exposure duration relationships. Using the animal lethality data of Barcroft (1931), ten Berge et al. (1986) calculated a mean value of 2.7 for n for six species of animals (range, 1.6 to 4.3). The value for the monkey was 1.9 and the value for the rat was 1.6. Using rat and mouse LC₅₀ data sets and exposure times of 5 to 60 min, Hilado and Cumming (1978) calculated an n value of 2. These data indicate a mean n value of 2. Additional data sets were available for the calculation of n values in the present document. Regression analysis of the rat lethality data by E.I. du Pont de Nemours (1981) for exposure durations of 5, 15, 30, and 60 min results in an n value of 2.6 (Appendix A, Figure A-2), and regression analysis of the rat lethality data of Ballantyne (1983), for the exposure durations of 5, 30, and 60 min, results in an n value of 2.1 (data not graphed).

FIGURE A-2

Regression Line for LC₅₀ values in rats (dta of E.I du Pont de Nemours [1981])

It should be noted that extrapolation of the rat 1-h LC₅₀ value of 139 ppm of E.I. du Pont de Nemours (1981) to 6 h (using C^2.6×t=k) results in a value of 70 ppm, which is similar to the rat LC₃₀ value of Blank (1983), 68 ppm, illustrated in Figure 5–1. Similar results from two different studies support the n value of 2.6 for extrapolation across time in lethality studies with the rat.

5.1. Human Data Relevant to AEGL-1

The odor threshold, 0.58 ppm to 5.0 ppm (Amoore and Hautala 1983; Ruth 1986) is low compared with irritant or toxic concentrations. No acute exposures were located resulting in mild effects in humans. Three monitoring studies, involving no symptoms to mild symptoms during chronic occupational exposures of adult males, are relevant to development of AEGL-1 values. The symptoms and blood concentrations of cyanide in the monitoring study of Chandra et al. (1980) indicate that the workers may have been exposed at higher atmospheric concentrations than those reported.

Mean concentrations of cyanide in the breathing zone of workers (all nonsmokers) in an electroplating area of three factories were 6, 8, and 10 ppm (range, 4.2–12.4 ppm) (El Ghawabi et al. 1975). Employment ranged from 5 to 15 y. Complaints of headache, weakness, and changes in taste and smell were reported by approximately 80% of the workers; incidences were much higher than in a matched control group. Irritation of the throat, vomiting, and effort dyspnea were commonly reported, and lachrimation and precordial pain were reported relatively less frequently. Two workers in the factory with the highest exposures suffered from psychotic episodes during the survey. Twenty of the 36 workers had thyroid enlargement to a mild or moderate degree. Air cyanide concentrations and exposure durations were not linked to specific symptoms. Mean levels of thiocyanates in the urine correlated with air concentrations of cyanide. Although the sample size was small, 36 workers, the study used 20 well-matched controls and a biological index of exposure (urinary thiocyanate). An NRC subcommittee concluded from this study that 1-h exposures at 8 ppm might produce mild headache in healthy adults (NRC 2000).

The Leeser et al. (1990) study was a controlled study with comprehensive medical examinations. In this study, presumably healthy workers were exposed to geometric mean HCN concentrations up to 1 ppm (range, 0.01–3.3 ppm) determined by personal monitoring in the work areas. Concentrations in the atmosphere of the plant ranged up to 6 ppm during the fall of the year, as indicated by Draeger pump tests or static monitors. It is not clear that the geometric mean concentrations include the later, higher values, as ranges during the spring were reported to be up to only 3.3 ppm. Higher blood cyanide levels were correlated with the higher exposure levels in the fall of the year. The results of clinical histories and medical examinations showed no differences to only minor differences for a variety of parameters between the HCN workers and a matched control group.

Medical questionnaires were not given in the Grabois (1954) study. However, both NIOSH (1976) and ACGIH (1996) reviewed the study. NIOSH (1976) identified 5 ppm as a no-effect concentration using the data for the five plants presented by Grabois (1954). Similar exposures were reported in the studies of Hardy et al. (1950) and Maehly and Swensson (1970).

5.2. Animal Data Relevant to AEGL-1

Animal studies that addressed sensory irritation or mild effects were not clearly distinguishable from those that addressed more severe effects.

5.3. Derivation of AEGL-1

The AEGL-1 is based on monitoring studies in which the preponderance of data as a weight-of-evidence consideration indicates that an 8-h exposure to 1 ppm would be without adverse effects for the general population. El Ghawabi et al. (1975) reported symptoms such as headache, weakness, changes in taste and smell, irritation of the throat, vomiting, and effort dyspnea in three electroplating plants in which mean concentrations of HCN were 6, 8, and 10 ppm, but the authors failed to relate symptoms to air concentrations. It should be noted that 20 of the 36 workers in the El Ghawabi et al. (1975) study had thyroid enlargement, which is characteristically observed in cases of chronic cyanide exposure and may have been responsible for some of the symptoms. An NRC subcommittee, in evaluating the El Ghawabi et al. (1975) data, concluded that the average concentration of 8 ppm in the three plants would likely produce no more than mild headache, which would be acceptable for a 1-h exposure of healthy adults. In the monitoring study of Leeser et al. (1990), chronic exposure of 63 workers in a cyanide salt production plant to geometric mean concentrations up to approximately 1 ppm and possible excursions up to 6 ppm during part of the year produced no clear exposure-related symptoms. According to NIOSH (1976), chronic exposure of workers to 5 ppm while processing apricot kernels in the monitoring study of Grabois (1954) was without effect. Additional monitoring studies with mean exposures to 5 ppm failed to report adverse health effects (Hardy et al. 1950; Maehly and Swensson 1970). It is unlikely that the population of workers in these and additional monitoring studies represent only healthy individuals.

The AEGL-1 was derived from a consideration of the dose-response data, which were obtained from all of the monitoring studies and subsequently time scaled to the shorter exposure durations. Although the exposures were of chronic duration in the monitoring studies, they represent the best available human data. Symptoms observed during chronic exposures should represent the greatest potential response. An 8-h exposure duration was selected as the basis for AEGL development.

Mild headache is a symptom of exposure that meets the definition of an AEGL-1. Dividing the 8-h concentration of 5 ppm of the Grabois (1954), Hardy et al. (1950), or Maehly and Swensson (1970) study by an intraspecies uncertainty factor (UF) of 3 or dividing the 1-h concentration of 8 ppm of the El Ghawabi et al. (1975) study by an intraspecies UF of 3 results in very similar AEGL-1 values. The resulting 8-h value of 1.7 ppm is also similar to the 8-h no-effect concentration of 1 ppm in the Leeser et al. (1990) study, where no UF was applied. UFs are generally applied to the highest NOAELs or lowest LOAELs. A UF was not applied to the Leeser et al. (1990) study because it was the lowest NOAEL. No specific susceptible populations were identified during numerous occupational monitoring studies or during the clinical use of nitroprusside solutions to control hypertension. Thus, potential differences in susceptibility among humans are not expected to exceed 3-fold. All individuals, including infants, possess large amounts of the cyanide detoxifying enzyme rhodanese (as well as other detoxifying enzymes) and normally have adequate amounts of sulfur-containing compounds.

The 8-h no-effect mean geometric concentration of 1 ppm (with excursions up to 6 ppm) from the Leeser et al. (1990) study was used as the basis for time scaling the AEGL-1 values. This study was chosen because it was well conducted: all workers had full medical examinations and routine blood tests, including measurements of blood cyanide and carboxyhemoglobin. Atmospheric HCN concentrations were monitored in the plant several times during the year. Because of the extrapolation from a long-term exposure, the 8-h value of 1 ppm was time scaled to the other exposure durations using the relationship C³×t=k where k=480 ppm³·min. In order to stay below the highest measured concentration from personal samplers, at 3.3 ppm in the Leeser et al. (1990) study, the 10-min value was set equal to the 30-min value. Calculations are in Appendix B, and values appear in Table 5–6 below.

TABLE 5–6

AEGL-1 Values for Hydrogen Cyanide.

6.1. Human Data Relevant to AEGL-2

As noted above for the AEGL-1, chronic occupational exposure of adult males to >10 ppm produced symptoms of headache, weakness, changes in taste and smell, irritation of the throat, vomiting, and effort dyspnea (El Ghawabi et al. 1975; NIOSH 1976; Blanc et al. 1985). For a few individuals, chronic exposures occasionally produced more serious adverse effects, such as fainting and psychotic episodes. There was no evidence that these symptoms occurred after one exposure. A concentration of ≥25 ppm for 1 h resulted in numbness, weakness, vertigo, nausea, rapid pulse, and flushing of the face (Parmenter 1926). Only one individual was involved, and neither the exposure duration nor the concentration were measured.

6.2. Animal Data Relevant to AEGL-2

Several animal studies listed in Table 5–5 describe effects at concentrations below those causing incapacitation or unconsciousness. These 30-min studies are as follows: monkey, 60 ppm (slight CNS effects) (Purser 1984); rat, 55 ppm (changes in lung dynamics and phospholipids) (Bhattacharya et al. 1984); and mouse, 63 ppm (respiratory depression of 50%) (Matijak-Schaper and Alarie 1982). From the description given by Matijak-Schaper and Alarie (1982), the concentration of 63 ppm for 30 min appears to be the threshold for a breathing pattern characteristic of asphyxiation. The effects in these three studies are reversible and do not impair the ability to escape, but they can be considered close to the threshold for such effects. Incapacitation in monkeys occurs at higher concentrations (80–150 ppm) (Purser et al. 1984). The 24-h exposure of mice at 30 ppm (Pryor et al. 1975), resulting in lung congestion, is also relevant to the definition of AEGL-2. The data of Sakurai (1989), incapacitation in mice inhaling 41.7 ppm in rotating cages for 30 min, appear low compared with the other studies and were not considered.

6.3. Derivation of AEGL-2

Because the human exposure concentrations are less reliable than the experimental animal data, the animal data were used in the derivation of the AEGL-2 values. The study chosen for the AEGL-2 derivation was the study by Purser (1984) with the monkey, because it was well conducted and used an appropriate species (compared with the rodent, the respiratory tracts of humans and monkey are more similar in anatomy, the amount and distribution of types of respiratory epithelia, and airflow pattern). This concentration was 60 ppm for 30 min. Although this end point (a slight depressive effect on the central nervous system as evidenced by a change in brain-wave activity near the end of the exposure) was a NOAEL for the definition of an AEGL-2, it was chosen because the next higher experimental concentration resulted in severe adverse effects of incapacitation, unconsciousness, and possibly death. The Barcroft (1931) lethality and incapacitation study has shown that the monkey is less sensitive to the respiratory and central nervous system effects of HCN than the rat and mouse (by factors of 1.75 and 3, respectively), and the adult human is less sensitive than the dog. The differences in sensitivity were based, at least partially, on the more rapid respiratory rates and greater cyanide uptake of rodents and the dog compared with humans and the monkey.

Because the respiratory tracts of humans and monkeys are more similar than those of humans and rodents, because uptake is more rapid in the monkey than in humans, and because both species have been shown to be relatively insensitive to the incapacitative and lethal effects of HCN (but at the same time, species susceptibilities to lethal effects do not differ by more than a factor of 1.5), an interspecies UF of 2 was applied. Human (adult) accidental and occupational exposures (El Ghawabi et al. 1975) indicate that there are individual differences in sensitivity to HCN, as evidenced by symptoms following chronic exposures, but the magnitude of these differences does not appear to be great. These studies and the clinical use of nitroprusside solutions to control hypertension do not demonstrate a susceptible population. The detoxifying enzyme rhodanese is functional in all individuals, including newborns. Therefore, a UF of 3 was applied to account for potential differences in human susceptibility. For the concentration-exposure duration relationship, the mean value for n of 2.0 for the monkey was calculated from two data sets involving incapacitation (2.1) and lethality (1.9) (Section 4.4.3). The 30-min exposure value of 60 ppm was divided by a total UF of 6 and scaled across time using the Cⁿ×t=k, where n=2 and k=3,000 ppm²·min. Values appear in Table 5–7 below, and calculations are in Appendix B.

TABLE 5–7

AEGL-2 Values for Hydrogen Cyanide.

The safety of the values is supported by the data of Grabois (1954), in which occupational exposures ranged up to 17 ppm, and two additional animal studies. The 30-min exposure of rats at 55 ppm (Bhattacharya et al. 1984), when divided by a total UF of 6 (2 for interspecies and 3 for intraspecies), results in a 30-min AEGL-2 of 9.2 ppm. The described effects of changes in lung dynamics and lung phospholipids are not irreversible or long-lasting. Mice experienced a decrease of 50% in respiratory rate when inhaling 63 ppm for 30 min but did not lose consciousness (Matijak-Schaper and Alarie 1982; Alarie 1997). Dividing by a total UF of 6 results in a 30-min AEGL-2 value of 11 ppm.

7.1. Human Data Relevant to AEGL-3

No human studies of sufficient exposure duration with measured concentrations producing irreversible or life-threatening effects were located in the available literature. However, the data of Barcroft (1931), a 1.5-min exposure at 500–625 ppm, and Bonsall (1984), a 6-min exposure at approximately 450 ppm, with recovery from symptoms and effects, can be considered short-term upper limits for healthy adults.

7.2. Animal Data Relevant to AEGL-3

LC₀₁ values for four time periods were provided by E.I. du Pont de Nemours (1981) for the rat. They are as follows: 5 min, 283 ppm; 15 min, 138 ppm; 30 min, 127 ppm; and 60 min, 88 ppm. Ballantyne (1983) used several concentrations and exposure durations but did not provide the actual concentrations; therefore, an LC₀₁ could not be calculated. Matijak-Schaper and Alarie (1982) reported no deaths in mice inhaling HCN at 100 ppm for 30 min. Mice inhaling HCN at 30 ppm for 24 h showed signs of lung congestion (Pryor et al. 1975).

7.3. Derivation of AEGL-3

The 15- and 30-min and 1-h LC₀₁ values of 138, 127, and 88 ppm, respectively, provided by E.I. du Pont de Nemours (1981) for the rat were used to derive the AEGL-3 values. Lethal concentrations are very similar for various animal species (Table 5–4), and Barcroft (1931) has shown that man and the monkey are less sensitive to the effects of HCN than are the rat and dog, a conclusion based at least partially on relative respiratory rates. Relative to body weight, humans have a much lower respiratory rate and cardiac output than rodents. These are the primary determinants of systemic uptake of volatile chemicals. Thus, at similar exposure concentrations, rodents will absorb substantially more cyanide than primates. Lower rhodanese activity levels in primates will not be significant at high, acute HCN exposure levels. These factors might argue for use of an interspecies UF of 1. However, an interspecies UF was applied because of the high acute toxicity and rapid action of HCN. Because LC₅₀ values among animal species differed by less than a factor of 2, an interspecies UF of 2 was applied. Human accidental and occupational exposures indicate that there are individual differences in sensitivity to HCN, but the magnitude of these differences does not appear to be great. No specific data on susceptible populations were located in numerous published monitoring studies or during the clinical use of nitroprusside solutions to control hypertension. The detoxifying enzyme rhodanese, as well as other enzymes, is functional in all individuals, including newborns. Therefore, a UF of 3 was applied to protect susceptible individuals. The concentration-exposure duration relationship for this data set is C^2.6×t=k (Section 4.4.3); therefore, the value of 2.6 for n was applied. The 15- and 30-min and 1-h values were divided by a total UF of 6 and the 15-min and 1-h values were scaled across time using the C^2.6×t=k relationship (the 15-min value for the 10-min AEGL-3 and the 1-h for the 4- and 8-h AEGL-3 values). Values appear in Table 5–8 and calculations are in Appendix B.

TABLE 5–8

AEGL-3 Values for Hydrogen Cyanide.

The AEGL values are supported by the study of Pryor et al. (1975) with the mouse in which a 24-h exposure at 30 ppm induced pulmonary congestion but was not lethal. The 30 ppm concentration divided by a total UF of 6 and scaled across time from 24 h to 30 min using C^2.6×t=k results in a 30-min AEGL-2 of 22 ppm. The AEGL values are also supported by the study of Parmenter (1926) in which an individual potentially exposed at 25–75 ppm for part of a day had severe symptoms but recovered fully. Furthermore, Barcroft's subject withstood a 1.5-min exposure at 500–625 ppm (Barcroft 1931). Time scaling the AEGL-3 values to 1.5 min results in a concentration at 60 ppm, which is less than the actual exposure by a factor of approximately 10.

8.1. AEGL Values and Toxicity End Points

The AEGL values and toxicity end points are summarized in Table 5–9.

TABLE 5–9

Summary of AEGL Values (ppm [mg/m³]).

8.2. Comparisons with Other Standards and Guidelines

Standards and guidance levels for workplace and community exposures are listed in Table 5–10. The Leeser et al. (1990) study was not available at the time many of these standards and guidelines were developed. The American Industrial Hygiene Association (AIHA 2000) did not derive an ERPG-1 value. The 1-h AEGL-2 and AEGL-3 values are slightly lower than the corresponding 1-h ERPG values. The ERPG-2 value was based on the Wexler et al. (1947) study in which sodium cyanide given intravenously to human volunteers at 0.11 mg/kg caused no deaths or serious injuries. The AIHA suggested that the intravenous dose approximates a 1-h exposure at 10 ppm. Because a bolus dose of cyanide does not take into account metabolism over the 1-h exposure duration, a well-conducted animal study was chosen with an appropriate species, but inter- and intraspecies UFs were applied. The ERPG-3 was based on several animal studies, including Purser (1984), in which concentrations of 45 to 60 ppm resulted in only reversible effects. These studies, and several additional lethality studies, were also reviewed for the AEGL-3.

TABLE 5–10

Extant Standards and Guidelines for Hydrogen Cyanide.

The 1-h AEGL-2 (7.1 ppm) is close to the 1-h Spacecraft Maximum Allowable Concentration (SMAC) of 8 ppm (NRC 2000), and both groups considered available monitoring studies in their derivations. Although the SMAC definition is similar to the AEGL-1 definition, the SMAC applies to healthy adults, whereas the AEGL-2 applies to the general population; therefore, the AEGL-2 value is conservative in comparison with the SMAC. The NRC subcommittee on SMACs used the monitoring data of El Ghawabi et al. (1975) to develop the values. The subcommittee suggested that the average concentration of “8.0 ppm in the three plants would likely produce no more than mild CNS effects (e.g., mild headache), which would be acceptable for 1-hour exposures in a spacecraft.” The subcommittee concluded that it was “likely that the more serious symptoms, such as vomiting, were the result of brief exposures to high HCN concentrations.” Therefore, 8 ppm was identified as the 1-h allowable concentration of HCN. The 24-h SMAC is 4 ppm and the 7-d SMAC is 1 ppm.

The NIOSH immediately dangerous to life and health (IDLH) value (NIOSH 1994) is greater than the 30-min AEGL-3. NIOSH based their recommended exposure limit (REL) on the statement by Flury and Zernik (1931) that 45–54 ppm could be tolerated by man for 0.5 to 1 h without immediate or late effects. Although the Flury and Zernik (1931) data are based on animal studies, NIOSH did not apply a UF.

Both ACGIH (1996) and NIOSH (1999) based their ceiling and short-term exposure limits, respectively, on one of the studies used for development of the AEGL-1. The value for both agency limits of 4.7 ppm was based on symptoms described during chronic exposures of workers in several studies and specifically on El Ghawabi et al. (1975). In 1993, the ACGIH value was reduced from 10 ppm to minimize the potential for irritation to the respiratory tract as well as potential acute and chronic effects of cyanide. The German and Dutch occupational exposure concentrations, analogous to the 8-h ACGIH time-weighted average (TWA), are 4.7 and 10 ppm, respectively. The German maximum workplace concentration (MAK) peak category or ceiling value is two times the MAK; this concentration is of a 5-min maximum duration and must not be exceeded at any time during the work shift.

8.3. Data Adequacy and Research Needs

The data base from animal studies is robust, but there are little definitive data on human exposure concentrations for short exposure durations and no definitive data on differences in susceptibilities among adults or between adults, infants, and children, other than well-understood ventilatory differences in the latter case. Gender and age-related differences in response to chronic cyanogenic glycoside consumption are difficult to interpret due to confounding, marked protein deficiencies in those populations that consume cassava as a major dietary staple (see Section 4.4.2). However, monitoring studies of presumably healthy adults that established no effect and/or minor discomfort concentrations to inhaled cyanide were available to set projected safe levels for the entire population by applying appropriate uncertainty factors (UFs). The metabolism and mechanism of action of cyanide are well understood and identical in all mammalian species. Data were available on concentrations involving lethal and sublethal effects for the monkey, dog, rat, mouse, and rabbit. Exposure durations included those ranging from a few seconds to 24 h. Where different mammalian species were tested in the same study, the results indicated that sensitivity to cyanide toxicity is similar among species, but slight differences may be related to body size, which in turn is related to respiration rate. Thus, establishing safe levels for humans based on small mammalian species adds confidence to the AEGL derivation. Animal studies with different toxicologic end points were available to establish concentration-exposure duration relationships. The extreme toxicity of HCN precludes certain types of tests, including long-term inhalation studies; therefore, genotoxicity, carcinogenicity and developmental and reproductive studies were performed with cyanide salts.

Several studies provided data on blood and urine concentrations of cyanide and thiocyanate following occupational exposures at low concentrations. These values are generally similar to those of smokers who have not been occupationally exposed to HCN. Whole-blood cyanide concentrations during nitroprusside infusion also have been measured and related to symptoms. There are also data on nonlethal oral doses and metabolism rates in humans. Taken together, the data indicate that the HCN AEGL values may be conservative. However, data on infants, children, and the elderly, populations that may be more susceptible to HCN toxicity than healthy adults based on higher respiration rates and slower metabolism, among other factors, are lacking. Furthermore, occupational monitoring data were collected under normal working conditions; stress or physical exertion may be greater under emergency conditions. Because HCN is extremely toxic and the range of human susceptibility is not definitively known, the AEGL derivations make use of appropriate UFs.

APPENDIX A. TIME-SCALING CALCULATIONS FOR HYDROGEN CYANIDE

APPENDIX B. DERIVATION OF AEGL VALUES

APPENDIX C. DERIVATION SUMMARY FOR ACUTE EXPOSURE GUIDELINE LEVELS FOR HYDROGEN CYANIDE (CAS No. 74–90–8)