Diagnosis and misdiagnosis of poisoning with the cyanide precursor acetonitrile: Nail polish remover or nail glue remover?

Diagnosis and Misdiagnosis of Poisoning With the Cyanide Precursor Acetonitrile: Nail Polish Remover or Nail Glue Remover? PETRIE

M.

RAINEY,

MD,

PHD,

WILLIAM

L. ROBERTS,

Accurate diagnosis of acetonitrile ingestion is critical to management. Often this involves differentiating nail polish remover (acetone) from nail glue remover (acetonitrlle). tnitial symptoms of acetonitrlle ingestion are indlstingulshable from those of acetone and common alcohols. However, acetonitrlle is metabolized to cyanide, producing severe delayed toxicity. Acetonitrile produced increased serum osmolallty and osmolal gap, but these flndlngs are non-specific and normal values cannot rule out potentially fatal exposure. Acetone, but not acetonitrlle, was detectable in urine or serum with Acetest tablets; both were unreactive with a ketone dipstick. Acetone and acetonitrile could be detected with routine gas chromatography methods for alcohols. Both substances had identical retentlon times on the widely used stationary phase, 5% Carbowax 20M on graphitlzed carbon, and with GasChrom 254. Three other systems afforded unique retention times, but acetonitrile was easily mistaken for ethanol In two. Physicians and laboratories must take care to avoid misdiagnosis of acetonitrile ingestion as exposure to acetone, ethanol or another alcohol. (Am J Emerg Med 1993;11:104-108. Copyright 0 1993 by W.B. Saunders Company)

Cosmetics and personal care products account for the highest frequency of pediatric ingestions reported to poison control centers. However, during the years 1983 to 1990, only two substances in this class caused fatalities: artificial

nail glue remover containing acetonitrile and mouthwash containing ethanol.’ The acetonitrile fatality resulted from an initial failure to appreciate that the ingestion involved nail glue remover, not nail polish remover.2 Nail polish removers consist principally of acetone. When ingested accidentally or intentionally, acetone is a mild central nervous system depressant and is eliminated without metabolism in the breath and urine. The amount of acetone in a typical bottle of nail polish remover is small and rarely causes significant poisoning. Artificial nail glue removers may consist primarily of aceton&rile. Several cases of accidental or intentional ingestion of these products have recently been reported.*-’ Acetonitrile is metabolized to cyanide in the body. After a latent period of several hours, symptoms of cyanide poisoning may ensue.* Proper diagnosis and management of acetonitrile ingestion before toxic cyanide concentrations are reached can prevent significant toxicity and result in full recovery. From the Department of Laboratory Medicine, Yale University School of Medicine, New Haven, CT. Manuscript received August 6, 1992; revision accepted September 22, 1992. Address reprint requests to Dr Rainey, Department of Laboratory Medicine, CB 506 Yale-New Haven Hospital, 20 York St, New Haven, CT 06504. Key Words: Acetone, acetonitrile, alcohol, cyanide, gas chromatography, poisoning. Copyright 0 1993 by W.B. Saunders Company 0?35-6757l93/1102-0002$5.00/O 104

MD,

PHD

Many clinical laboratories offer an alcohol or volatiles screen by gas chromatography that usually includes identification and quantitation of ethanol, methanol, isopropanol and acetone. These tests might be used to confirm acetone ingestion. They also should be able to detect acetonitrile. An elevated osmolal gap also may result from either acetonitrile or acetone ingestion. The finding of a positive reaction for ketone bodies in either the serum or urine would indicate a presumptive diagnosis of acetone ingestion. The latter tests can be helpful in situations where gas chromatography is unavailable. We investigated the application of gas chromatography, osmolality and ketone body measurements to the diagnosis of acetonitrile ingestion. All are useful and gas chromatography can be definitive. However, osmolality cannot distinguish the presence of acetonitrile from other more common solvents, including acetone. Some dipsticks are insensitive to acetone and fasting ketosis may give a false positive. Positive chromatographic findings could possibly be misinterpreted as indicating the presence of acetone or ethanol. The failure to recognize an acetonitrile ingestion could result in a potentially fatal misdiagnosis. METHODS Specimens Pooled, lipid-stripped human serum (Scantibodies Laboratory, Inc, Santee, CA) was spiked with acetonitrile, methanol, ethanol, acetone and/or isopropanol (all high-performance liquid chromatography or reagent grade). Drug-free normal human urine was spiked with acetonitrile or acetone.

Osmolal Gap Serum specimens spiked with increasing concentrations of acetonitrile were analyzed in triplicate for sodium, urea nitrogen, and glucose on an Astra 8 analyzer (Beckman Instruments, Inc, Brea, CA) and for osmolality on a One-Ten osmometer (Fiske Associates, Needham, MA). Osmolal gap was calculated as A osmolality = osmolality (mOsm/kg) - 2 x Na (mmol/L) + glucose (mg/dL)/lS + urea nitrogen (mg/dL)/2.8.

Ketone Bodies Spiked serum and urine specimens were tested with the Ames Multistix dipstick and Acetest Tablets (both from Miles, Inc, Elkhart, IN) according to the manufacturer’s instructions. Color was compared directly with negative controls.

Gas Chmmatography A lOO+L sample of the specimen was mixed with 500 p,L of water containing 0.5 mL/L of n-propanol as an internal standard and 0.1% sodium fluoride as a preservative. Aliquots of 1.O )LL of the resulting mixture were manually injected directly into the inlet port of a gas chromatograph. Retention times (the characteristic time between

RAINEY AND ROBERTS

n ACETONITRILE

POISONING

injection and detection of the analyte peak) were determined for each analyte in each system. System 1. A Sigma 4 gas chromatograph (Perkin-Elmer Corp, Norwalk, CT) with a flame ionization detector was fitted with a 6’ x YE”stainless steel column packed with 5% Carbowax 20M on acidwashed 80 to 120 mesh Graphpac-GB (Alltech, Deertield, IL). Oven temperature, 85°C; injector, 175°C; detector, 175°C; carrier, helium at 30 ml/mitt. System 2. A Perkin-Elmer 8410 gas chromatograph with a flame ionization detector was fitted with an 18” x l/g”stainless steel column packed with 2% Carbowax 20M on acid-washed 80 to 120 mesh Graphpac-GB (Alltech). Oven temperature, 75°C; injector, 130°C; detector, 120°C; carrier, helium at 30 mlimin. System 3. As in System 1, but using a packing of 0.2% Carbowax 1500 on 80 to 100 mesh Carbopack-C (Alltech) and an oven temperature of 100°C. System 4. As in System 1, but using a packing of 80 to 100 mesh GasChrom 254 (Alltech) and an oven temperature of 140°C. Acetone and isopropanol comigrated in this system under a wide variety of conditions. System 5. As in System 1, but using a packing of 10% poly(diethyleneglycol succinate), phosphoric acid stabilized (DEGS-PS) on 80 to 100 mesh Supelcoport (Supelco, Inc. Bellefonte, PA) and an oven temperature of 70°C.

Octanol-Water Partition Coefficients Solutions of acetonitrile, ethanol and methanol were prepared at 0.1 volume percent and 2.0 mL aliquots were extracted with 2.0 mL portions of I-octanol (high-performance liquid chromatography grade; Aldrich, Milwaukee. WI) for 30 minutes in a rotary mixer. After centrifugation for 5 minutes, the aqueous phases were removed and analyzed by gas chromatography on System 2. Controls consisted of 4 mL of the aqueous solutions carried through a sham partition. Concentration in the octanol phase was calculated as the difference between the concentration in the extracted phase and the control.

RESULTS Addition

of varying

concentrations

105

DIAGNOSIS

of acetonitrile

to

pooled human serum resulted in an increase in the serum osmolality and the osmolar gap of 1.05 mOsm/kg for each mmol/L increase in acetonitrile concentration, compared with a predicted increase of 1.08 mOsm/kg (based on a serum water content of 93%). As seen in Figure 1, serum osmolality did not exceed its upper limit of normal of 295 mOsm/kg until an acetonitrile concentration of 21 .O mmol/L (86 mg/dL) was reached, while the osmolal gap exceeded its upper limit of 10 mOsm/kg at an acetonitrile concentration of 12.4 mmoVL (51 mg/dL). The Multistix dipstick showed no reactivity with acetone or acetonitrile in either serum or urine at concentrations up to 100 mg/dL. Acetest tablets produced trace positive reactions when acetone was present in urine at concentrations above 10 mgldL and in serum at concentrations above 20 mg/dL. No reactivity was noted with acetonitrile in urine or serum at concentrations up to 100 mg/dL. In System 1 with a 5% Carbowax 20M stationary phase on a graphitized carbon support, acetonitrile had a retention time of 2.46 minutes. This was identical to the retention time for acetone in this system. Figure 2A shows the chromatography of the volatiles usually detected in this system (eg, methanol, ethanol, acetone and isopropanol), as well as the internal standard, n-propanol. Figure 2B shows the chromatogram of the same specimen to which acetonitrile had

Acetonitrile 0 504

Concentration

10

’

’

(mM)

20

30

40

’

’

’

325

40-

-315

3a-

-305

XI-

-295

IO-

- 285

0;

- 275

-10 0

50 Acetonitrile

100 Concentration

150

1265 200

(mg/dL)

FIGURE 1. Changes in serum osmolality produced by acetonitrile. Increasing amounts of acetonitrile were added to pooled human serum and the osmolal gap determined using the formula A Osmolality = Measured osmolality - (2 x sodium + glucose/l8 + BUN/2.8). The slope of the best fit line was 1.05 mOsm/kg for each mmol/L of acetonitrile.

been added. System 1 is very similar to one recommended in a major text of clinical chemistry6 and therefore likely to be widely used. The slight differences in the two methods are unlikely to affect the relative retention times. Several other stationary phases were also investigated to determine whether acetonitrile could be clearly identified. The results are summarized in Table 1. Carbowaxes are the most popular stationary phases for alcohol analysis and a total of three variations were investigated. Two other stationary phases were also investigated. The GasChrom 254 packing does not have a liquid stationary phase; it relies on adsorption to the porous polymer, rather than solution in a liquid coating. The DEGS-PS stationary phase is more polar than the Carbowaxes. Like the 5% Carbowax 20M, the GasChrom packing gave the same retention time for acetone and acetonitrile. Acetonitrile had unique retention times on the other Carbowax columns, but did not achieve baseline resolution from ethanol. On the polar DEGS-PS column, acetonitrile gave a distinctive peak with a retention time greater than the internal standard. Resolution of the other alcohols on the DEGS-PS column was less than optimal. The octanol-water partition coefficients were 0.15 for methanol, 0.18 for acetonitrile, and 0.42 for ethanol. DISCUSSION The availability of acetonitrile to the public in the form of glue remover for artificial nails has led to both accidental and intentional ingestions. Several cases have been reported recently and have included reviews of the somewhat limited literature.2-5 It may not always be clear whether nail glue remover (acetonitrile) or nail polish remover (acetone) is involved. The initial symptoms of both ingestions are similar. The toxicity of acetone remains low, but acetonitrile is me-

AMERICAN

106

JOURNAL

5

01234567

Time, min FIGURE 2. Volatiles analysis by gas chromatography using 5% Carbowax 20M on Graphpac-GB. Specimens were prepared and analyzed as described under Methods using System 1. (Top) Serum containing 50 mg/dL of methanol and 150 mg/dL each of ethanol, isopropanol and acetone. (Bottom) Serum used in the top panel to which 100 mg/dL of acetonitrile had been added. tabolized slowly to cyanide. Severe poisoning can result after a latent period of 3 to 12 hours.* As little as 5 mL may be fatal in a small child.5 Failure to distinguish between ingestion of nail polish remover and nail glue remover has resulted in at least one fatality.* Accurate early diagnosis of acetonitrile ingestion should

OF EMERGENCY

MEDICINE

n Volume

11, Number 2 W March 1993

result in appropriate treatment and complete recovery with no sequelae. The body has a high capacity to metabolize cyanide to the much less toxic thiocyanate ion. The rate of metabolism is limited primarily by the availability of sulfur. Infusions of thiosulfate have produced rapid lowering of cyanide levels in patients poisoned with acetonitrile.3*7 Hemodialysis was also very effective in reversing toxicity in one case. Presumably both acetonitrile and cyanide were effectively removed (as well as thiocyanate from previous thiosulfate treatment); however, levels were not measured.3 Nitrites and hydroxycobalamin are useful in the treatment of an acute cyanide overdose, but may have less utility in the face of continuous cyanide production. The osmolal gap has been a traditional test for ingestion of low molecular weight water-miscible solvents, such as ethanol, methanol, isopropanol, ethylene glycol and acetone. Ingestion of acetonitrile should produce an increase in osmolality slightly greater than that produced by a similar amount of ethanol. This is predicted from its slightly lower molecular weight (41, v 46 for ethanol) and an expectation of similar volumes of distribution. The volume of distribution for acetonitrile has not been determined in humans, but can be extrapolated from the behavior of ethanol and methanol, both of which distribute into total body water (Vd = 0.5 to 0.6 L/kg).’ Like these two substances, acetonitrile is 100% miscible with water. Its lipophilicity, as measured by its octanol-water partition coefficient of 0.18, lies between those of methanol and ethanol (0.15 and 0.42, respectively). Thus, acetonitrile also should distribute into total body water. A peak acetonitrile level was measured in one case of ingestion and was consistent with this prediction.7 Figure 1 shows that acetonitrile produced changes in osmolality very close to those predicted from the amount added. A problem with using osmolality-based measurements is that no distinction can be made between acetonitrile and other more common substances, including acetone. Like methanol and ethylene glycol, acetonitrile is a toxic solvent that may produce an elevation in the osmolality or osmolal gap. However, normal values for these parameters may also be found in association with toxic ingestions of any of these substances. It is not widely appreciated that a normal osmola1 gap is often negative.’ In Figure 1 the acetonitrile concentration had to be above 12.4 mmol/L (51 mg/dL) for the osmolal gap to exceed the upper limit of normal of 10 mOsm/kg. Serum osmolality was less sensitive, requiring an acetonitrile concentration of 2 1.O mmol/L (86 mg/dL) to exceed the upper limit of 295 mOsm/kg (This represents an extreme case, since the starting osmolality was only 273 mOsm/kg, a value that would be exceeded in 99% of a healthy population). These levels correspond to ingestions of about 0.3 g/kg and 0.5 g/kg, either of which is a potentially fatal dose. Acetonitrile should now be included in the differential diagnosis of a systemic acidosis with both anion and osmolar gaps. Like methanol and ethylene glycol, it has a low intrinsic toxicity, but yields a highly toxic metabolite that produces a delayed anion gap acidosis. It differs from the other two in that a lactic acidosis will be produced. Another distinctive feature expected in symptomatic acetonitrile poisoning is a reduced arteriovenous oxygen gradient, resulting from the inhibition of oxygen utilization by cyanide. Unlike

RAINEY AND ROBERTS

TABLE 1.

n ACETONITRILE

Retention Times of Acetonitrile

POISONING

107

DIAGNOSIS

and Other Voiatiles

Using Various Gas Chromatography Retention

Stationary

Phase

Carbowax 20M 0.2% Carbowax 1500 2% Carbowax 20M (18”) GasChrom 254 DEGS-PS

5%

Time (min)

Acetonitrile

Methanol

Ethanol

Acetone

2.46 1.55 0.56 2.25 1.75

1.42 1.08 0.28 0.85 1 .Ol

2.00 1.68 0.66 1.45 0.95

2.46 2.30 0.95 2.25 0.75

methanol and ethylene glycol poisoning, ethanol is not an antidote. Acetonitrile is metabolized to cyanide via cytochrome P-450,‘O not via alcohol dehydrogenase. It may be possible to differentiate ingestion of acetonitrile from acetone using tests for ketone bodies. These tests have high sensitivity for acetoacetate, but may also react with acetone at several-fold higher concentrations. The widely used Ketostix and Multistix dipsticks (Miles, Inc) use the same reagent and are insensitive to acetone according to the package inserts. We verified the lack of reactivity of the Multistix dipstick to acetone at 100 mg/dl. Some other urine strips may give positive reactions with acetone at high concentrations (eg. 50 mg/dl for the Rapignost test strip [Behring Diagnostics, Inc, Somerville, NJ] or 70 mg/dl for the Chemstrip [Boehringer Mannheim Corp. Indianapolis. IN], according to the manufacturers). Acetest tablets offered better sensitivity. Trace positive reactions were noted when acetone concentration exceeded 10 mg/dL in urine and 20 mg/dL in serum. No reactivity was seen with acetonitrile at concentrations up to 100 mg/dL in either matrix. Based on a volume of distribution of 0.8 L/kg’, ingestion of 1 mL/kg (0.8 g/kg) of acetone should result in a peak level of approximately 100 mg/dL. Thus, a trace positive urine reaction may be noted with acetone ingestions as low as 0.1 mL/kg (because acetone is a small, neutral molecule, it is readily resorbed in the renal tubule and is only slightly concentrated in urine relative to serum). It should be noted that children are more susceptible than adults to development of a fasting ketosis, which can lead to a false positive reaction. Because acetoacetate is concentrated in the urine, false positives are more likely if urine is used for testing. Gas chromatography offers the possibility of a definitive diagnosis. Many clinical laboratories offer screens for alcohols or volatile solvents by gas chromatography. These methods are commonly used for methanol, ethanol, acetone, and isopropanol. We have shown that such methods also can be used to detect and quantify acetonitrile. Acetonitrile will produce a peak in such systems whether or not it is being specifically sought. In several of the systems investigated, this peak occurred at retention times similar or identical to one of the more commonly seen analytes. Because few laboratories have experience with the analysis of acetonitrile, such a peak could be easily mistaken for the more commonly seen analyte. An acetonitrile-containing specimen submitted with a request for acetone analysis and analyzed by System 1 or System 4 would probably be reported as showing the presence of acetone, thereby apparently confirming the ingestion of relatively innocuous nail polish remover. With other methods, either a technologist or

Systems

lsopropanol 3.45 2.83 1.39 2.25 0.90

n-Propanol

4.80 3.71 2.08 3.19 1.47

an automated integrator might mistake the acetonitrile peak for whichever of the usual analytes had the closest retention time. If methods similar to System 2 or System 3 were used, a positive report for ethanol would be likely. Although careful measurement of the retention time could reveal the difference, past experience would offer no basis for expecting anything other than the usual analytes. The early symptoms of intoxication with all these agents are similar and would offer no clinical grounds for doubting an erroneous report. Misidentification is more likely with manual injection, since variations in injection technique can result in added variability in the retention time. The DEGS-PS column provided a very distinctive retention time for acetonitrile and can be recommended as a confirmatory method. This column can also be used for the measurement of ethylene glycol.” Unfortunately, conditions could not be devised which gave good resolution of methanol, ethanol, and isopropanol. We did not investigate the behavior of acetonitrile using capillary gas chromatography, since capillary columns are more susceptible to the effects of contamination inherent in rapid. direct injection methods suitable for analysis of emergency room specimens. We also did not investigate detection after head space sampling (the optimal technique for capillary columns) because the need for equilibration substantially lengthens turnaround time. Capillary systems have higher efficiency and resolution than packed columns, but no greater selectivity. Although baseline resolution of acetonitrile could be expected with a capillary column, the retention time may still be quite similar to that of a more common volatile, and the misidentification of a single isolated peak remains possible. GONCLUSIONS Nail glue remover contains acetonitrile, which is metabolized to cyanide. The early symptoms of intoxication produced by unmetabolized acetonitrile cannot be clinically distinguished from the symptoms produced by acetone, the principal ingredient in nail polish remover, or by other water-miscible solvents, including ethanol. Laboratory confirmation of the substance ingested can greatly assist in the correct management. Although few (if any) hospital laboratories include acetonitrile in their official list of tests, ingestion may be detected using osmolality measurements and differentiated specifically from acetone by the latter’s reactivity with Acetest tablets. Measurements of serum osmolality or osmolal gap have limited sensitivity and specificity, however, and Acetest tablets can yield a positive result from fasting ketosis. If available, gas chromatography offers the possibility of a

108

AMERICAN JOURNAL OF EMERGENCY MEDICINE I Volume 11, Number 2 n March 1993

definitive result. To avoid possible misidentification of acetonitrile as another more frequently seen volatile, laboratories should investigate its behavior in their assay systems. Specific inquiries about the possible presence of acetonitrile should be made on any specimen received for acetone analysis. Physicians should likewise specifically notify the laboratory of any specimen that might contain acetonitrile. As is true for many poisonings, optimum management requires good communication between the physician and the toxicology laboratory. REFERENCES 1. Litovitz T, Manoguerra A: Comparison of pediatric poisoning hazards: An analysis of 3.8 million exposure incidents. Pediatrics 1992;89:999-1006 2. Caravati EM, Litovitz TL: Pediatric cyanide intoxication and death from an acetonitrile containing cosmetic. JAMA 1988; 260~3470-3473 3. Turchen SG, Manoguerra AS, Whitney C: Severe cyanide poisoning from the ingestion of an acetonitrile-containing cosmetic. Am J Emerg Med 1991;9:264-267

4. Geller RJ, Ekins BR, lknoian RC: Cyanide toxicity from acetonitrile-containing false nail remover. Am J Emerg Med 1991;9:268-270 5. Kurt TL, Day LC, Reed WG, et al: Cyanide poisoning from glue-on nail remover. Am J Emerg Med 1991;9:271-272 6. Blanke RV, Decker WJ: Analysis of toxic substances. In Tietz NW (ed): Textbook of Clinical Chemistry. Chicago, IL, Saunders, 1986, pp 1670-1744 7. Michaelis HC, Clemens C, Kijewski H, et al: Acetonitrile serum concentrations and cyanide blood levels in a case of suicidal oral acetonitrile ingestion. J Toxicol Clin Toxicol 1991;29: 447-458 8. Baselt RC: Disposition of Toxic Drugs and Chemicals in Man (ed 2). Davis, CA, Biomedical Publications, 1982, pp 299, 491 9. Glasser L, Sternglanz PD, Combie J, et al: Serum osmolality and its applicability to drug overdose. Am J Clin Pathol 1973;60:695-699 10. Freeman JJ, Hayes EP: Microsomal metabolism of acetonitrile to cyanide. Biochem Pharmacol 1988;37:1153-1159 11. Cummings KC, Jatlow PI: Sample preparation by ultrafiltration for direct gas chromatographic analysis of ethylene glycol in plasma. J Anal Toxicol 1982;6:324-326