Free Radical and Drug Oxidation Products in an Intensive Care Unit Sedative: Propofol with Sulfite*

Free radical and drug oxidation products in an intensive care unit sedative: Propofol with sulﬁte*

Max T. Baker, PhD; Marc S. Gregerson, BS; Sean M. Martin, BA; Garry R. Buettner, PhD

Objectives: Some propofol emulsion formulations contain EDTA The EDTA propofol emulsion remained white at all times. The or sodium metabisulfite to inhibit microbe growth on extrinsic absorbance spectra of the propofol dimer and propofol dimer contamination. EDTA is not known to react with propofol formu- quinone extracted from sulfite propofol emulsion showed that lation components; however, sulfite has been shown to support propofol dimer did not absorb in the visible spectrum, but the some oxidation processes and may react with propofol. This study propofol dimer quinone had an absorbance peak at 421 nm, compared the oxidation of propofol and the formation of free causing it to appear yellow. Electron paramagnetic resonance radicals by electron paramagnetic resonance analysis in EDTA analysis of the propofol emulsion containing metabisulfite and sulfite propofol emulsions during a simulated intensive care revealed that the sulfite propofol emulsion yielded a strong unit 12-hr intravenous infusion. free radical signal consistent with the formation of the sulfite ⅐؊ Design: Controlled laboratory study. anion radical (SO3 ). The EDTA propofol emulsion yielded no free Setting: University laboratory. radical signal above background. Measurements and Main Results: Propofol emulsions (3.5 mL) Conclusion: Sulfite from the metabisulfite additive in propofol were dripped from spiked 50-mL vials at each hour for 12 hrs. emulsion creates an oxidative environment when these emul- Two propofol oxidation products, identified as propofol dimer and sions are exposed to air during a simulated intravenous infu- propofol dimer quinone, were detected in sulfite and EDTA propo- sion. This oxidation results in propofol dimerization and emul- fol emulsions; however, sulfite propofol emulsion contained sion yellowing, the latter of which is caused by the formation higher quantities of both compounds. After initiation of the sim- of propofol dimer quinone. These processes can be attributed ulated infusion, the quantities of propofol dimer and propofol to the rapid formation of the reactive sulfite free radical. (Crit dimer quinone increased in the sulfite propofol emulsion, but the Care Med 2003; 31:787–792) lower levels in the EDTA propofol emulsion remained constant. KEY WORDS: sulfite; propofol; metabisulfite; propofol oxidation; Sulfite propofol emulsion began to visibly yellow at about 6–7 hrs. free radicals; EDTA; propofol dimer; propofol dimer quinone

ropofol (2,6-diisopropylphe- lations are oil-in-H2O emulsions consist- analysis of the sulfite propofol emulsion nol) is an intravenous sedative ing of propofol (1%, 10 mg/mL), soybean and propofol emulsion containing EDTA that is administered continu- oil (100 mg/mL), egg yolk lecithin (12 (EDTA propofol emulsion) showed that ously to a wide variety of crit- mg/mL), and glycerol (22.5 mg/mL) in the sulfite propofol emulsion, but not the icallyP ill patients, including those having H2O (1). Because these emulsions can EDTA propofol emulsion, contained vari- coronary artery bypass grafts, adult respi- support the growth of bacteria and yeast able quantities of the lipid peroxidation ratory distress syndrome, asthma, head when extrinsically contaminated (2), an product, malondialdehyde, which in- trauma, status epilepticus, pneumonia, agent to retard the growth of microor- creased over time after air exposure (11). and sepsis (according to manufacturer ganisms is deemed necessary. Sodium Propofol dimer is a product of a one- prescribing information from Astra- metabisulfite (Na2S2O5, 0.25 mg/mL) (3, electron oxidation of propofol to form Zeneca Pharmaceuticals and Gensia Sicor 4) or EDTA (0.05 mg/mL) (5) are added to propofol radicals that couple with one Pharmaceuticals). Propofol is insoluble some propofol emulsions for that pur- another, generating the dimer product in aqueous media (propofol octanol/H2O pose. (6). Lipid peroxidation is a free radical ϭ partition coefficient 6800:1) and there- Drug additives are usually considered process involving the formation of lipid fore is formulated in a lipid-based vehicle. inactive formulation components; how- radicals in the presence of oxygen (12). It The current commercial propofol formu- ever, there is evidence that sulfite is re- is proposed that these sulfite-promoted sponsible for promoting the oxidation of chemical processes are initiated by the propofol and lipids in emulsions. For ex- formation of sulfite-derived radicals. *See also p. 981. ample, when propofol emulsions contain- The precise conditions for sulfite- From the Department of Anesthesia (MTB, MSG), ing metabisulfite (sulfite propofol emul- supported oxidation of propofol in sulfite ESR Facility (SMM), and Free Radical and Radiation sion) are exposed to air, a propofol propofol emulsions remain unclear. Fur- Biology (GRB), University of Iowa, Iowa City, IA. Supported, in part, by AstraZeneca Pharmaceuti- dimerized product appears (6). A yellow thermore, the detection of free radicals in cals LP, Wilmington, DE. coloration is also sometimes noted after emulsions of propofol has not been exam- Copyright © 2003 by Lippincott Williams & Wilkins exposure of the sulfite propofol emulsion ined. Twelve hours is the recommended DOI: 10.1097/01.CCM.0000053560.05156.73 to air (7–10). In addition, a comparative hang time for propofol emulsion in in-

Crit Care Med 2003 Vol. 31, No. 3 787 mins and an isocratic period of 100% methanol for 4 mins. A sample size of 5 ␮L was injected, and fractions in the eluate were scanned from 200 to 500 nm. The data were collected using the Xcalibur Software system (Herndon, VA). Propofol dimer was quanti- tated by its absorbance at 266 nm and propofol dimer quinone by its absorbance at 421 nm. Standard curves were constructed using standard propofol dimer and propofol dimer quinone. Electron Paramagnetic Resonance Spec- troscopy. Electron paramagnetic resonance (EPR) spectra were recorded with a Bruker EMX EPR spectrometer (Bruker Instruments, Billerica, MA) at room temperature. Samples were prepared by adding 5,5-dimethyl-1- pyrroline-N-oxide (DMPO, final concentration of 25 mM) to the propofol emulsions (total volume of 0.5 mL) from commercially supplied vials. EPR spectra were collected imme- diately after sample preparation. The emulsions were maintained at room temperature at all times. All spectra were collected using identical conditions: each represents the signal-averaged result of five scans; scan rate, 80 G/84 s; microwave power, 40 mW with a TM cavity and 10 mm flat cell; frequency, 9.77 GHz; modulation amplitude, 1.00 G; time constant, 82 ms. Figure 1. Liquid chromatography/mass spectral (MS) of propofol dimer quinone and secondary Values are reported as mean of duplicate or fragmentation (MS/MS) of the mass-to-change ratio (m/z) 353 ion. triplicate determinations (ϮSD). Statistics were performed by analysis of variance re- peated measures (Scheffe’s F test). p Values of Յ tensive care unit sedation, after which it out and collected for analysis. At each hour for .05 were considered significant. is recommended that the emulsion be up to 12 hrs, an additional 3.5-mL aliquot was discarded after vial spiking. The present released, collected, and analyzed. RESULTS study compared propofol degradation and Propofol Product Analysis. One milliliter ␮ free radical formation in EDTA and sulfite of collected emulsion was treated with 100 L High-pressure liquid chromatography propofol emulsions during a simulated of a 10% NaCl solution to crack the emulsion. analysis of ethyl acetate extracts of propo- intensive care unit 12-hr intravenous in- Ethyl acetate (1 mL) was added to the 1-mL fol emulsions revealed the presence of cracked emulsion, and the mixture was vigor- fusion in the laboratory. two major fractions that could be identi- ously shaken, followed by centrifugation to fied as propofol oxidation products. The ␮ separate the phases. Approximately 100 Lof first fraction chromatographed at approx- MATERIALS AND METHODS the ethyl acetate phase was taken for liquid imately 6.3–6.5 mins in the system used chromatography/mass spectral analysis. Chemicals. Fifty-milliliter vials of EDTA Liquid Chromatography/Mass Spectral and was identified as propofol dimer. The propofol emulsion and authentic propofol Analysis of Propofol Products. The ethyl ace- absorbance spectrum from 200 to 500 nm dimer and propofol dimer quinone were ob- tate extracts from propofol emulsions were of this peak showed an absorbance peak tained from AstraZeneca Pharmaceuticals analyzed on a Surveyor liquid chromatogra- at 266 nm and no absorbance in the vis- (Wilmington, DE). The propofol dimer and ible range. This compound chromato- propofol dimer quinone standards had purities phy/mass spectral analysis system (LCQDeca of Ͼ97%. Propofol emulsion containing met- MS system, ThermoFinnigan, San Jose, CA) graphed identically with authentic propo- abisulfite, also in 50-mL vials, was purchased having a photodiode array detector linked in fol dimer, and it exhibited the same from Gensia Sicor Pharmaceuticals (Irvine, series to the mass spectrum detector. The absorbance spectrum as authentic propo- CA). mass spectrum detector utilized atmospheric fol dimer, confirming that the propofol Simulated Intravenous Infusion. Vials of pressure chemical ionization and a normal- dimer is a visibly colorless compound. propofol emulsion were dripped by inserting a ized collision energy of 40%. This instrument The electrospray mass spectrum of the Flu-Vent vented spike and tubing (Flu Ven was equipped with a Discovery HS C-18 MS ϫ ␮ propofol dimer confirms the presence of 0153-C, Venusa, El Paso, TX) into intact, prop- analytical column (75 2.1 mm ID, 3 m) propofol dimer in these propofol emul- fitted with a guard column. Propofol and erly stored, commercially supplied vials of sions. Propofol dimer (molecular weight propofol product separation involved the use propofol emulsion. The vials were inverted ϭ 354 Da) poorly ionizes in electrospray, and hung on a ringstand at room temperature of a mobile phase gradient system. The initial under ambient fluorescent lighting. The tub- mobile phase was methanol/5 mM ammonium as does propofol, due to the stability of acetate (70/30, v/v) that was run at a flow rate the hydroxyl groups to ionization. How- ing from the vented spikes was cut to 30 cm. ϩ Emulsion release was regulated using the of 250 ␮L/min for 3 mins. This was followed ever, some M (mass-to-change ratio ϩ roller valve attached to the spike tubing. At by an increase in flow rate to 350 ␮L/min and [m/z] 354) and MH (m/z 355) ions are ϩ zero time, 3.5 mL of emulsion were dripped a linear gradient to methanol (100%) over 2 seen, as is the diagnostic m/z 313 (MH

788 Crit Care Med 2003 Vol. 31, No. 3 Ϫ ϩ Ϫ C3H6) ion. These ionization patterns quinone and had a similar absorbance 353 yielded 311 (MH C3H6), 269 ϩ Ϫ Ϫ ϩ also reflect the fact that propofol dimer spectrum as the standard. The standard (MH C3H6, C3H6), and 227 (MH ϩ Ϫ Ϫ Ϫ poorly ionizes with the addition of H propofol dimer quinone appeared pur- C3H6, C3H6, C3H6), representing (proton). This fragmentation pattern is ple in crystalline form but was yellow in cleavages of multiple isopropyl groups identical to the standard propofol dimer solution. during ionization. This mass spectrum is (not shown). Mass spectral analysis further con- identical to the mass spectrum of the The second fraction from the propo- firmed the identity of the propofol dimer standard propofol dimer quinone (not fol emulsions (retention time, 7.9– 8.1 quinone (Fig. 1), a previously unreported shown). mins) exhibited absorption in the ultra- compound. The propofol dimer quinone The effects of emulsion aeration on ϭ violet and visible range, having a major (molecular weight 352 Da) ionized ad- dripping on the quantities of propofol peak absorbance at 228 nm and a lesser equately for analysis with the addition of ϩ ϩ dimer in sulfite propofol emulsion and peak absorbance of 421 nm. This prod- H , yielding fragments of m/z 353 (MH ) EDTA propofol emulsion are shown in ϩ Ϫ uct chromatographed identically with and m/z 311 (MH C3H6). Secondary Figure 2. Initial propofol dimer levels the compound called propofol dimer fragmentation (MS/MS) of the ion m/z (time 0) were greater in sulfite propofol emulsion than in the EDTA propofol emulsion, and the concentrations increased in the former up to the 12 hrs examined. The low concentrations of propofol dimer in EDTA propofol emulsion did not increase over time. The formation of propofol dimer quinone in sulfite propofol emulsion and EDTA propofol emulsion during the simulated infusion is shown in Figure 3. Ini- tial propofol dimer quinone concentrations in sulfite propofol emulsion were also higher than in EDTA propofol emulsion and, like the propofol dimer, increased over time. The greatest rates of increase occurred from 6 hrs onward. The sulfite propofol emulsion visibly yellowed slightly, beginning at about 6–7 Figure 2. Propofol dimer in sulfite propofol emulsion and EDTA propofol emulsion during a simulated hrs. The propofol dimer quinone in EDTA 12-hr infusion. Values represent the mean (ϩSD) of triplicate determinations from three 50-mL vials propofol emulsion did not increase over of each propofol formulation. Sulfite propofol emulsions used were: one vial each of lots O0N301 time, and the emulsion remained visibly (expiration date, November 2002), O1A317 (expiration date, January 2003), and O1K310 (expiration white at all times. date, August 2003). EDTA propofol emulsions used were two vials of lot 4017F (expiration date, June Some of the yellowed sulfite propofol 2003) and one vial of lot 4359B (expiration date, July 2001). *Values were significantly different from corresponding 0-hr values. emulsion (1 mL) was treated with the chemical reductant sodium dithionite (1 mg) and examined by high-pressure liquid chromatography. Dithionite caused a bleaching of the yellow color, a decrease in propofol dimer quinone, and an increase in propofol dimer (data not shown). EPR Analysis. Figure 4 shows the EPR spectra obtained from the two different propofol formulations on exposure to air. The propofol emulsion that contains metabisulfite yields an EPR spectrum consistent with the spin trapping by DMPO of ⅐Ϫ H ϭ the sulfite-radical anion (SO3 ): a 16.0 G, aN ϭ 14.5 G (13, 14). The propofol emulsion containing EDTA rather than metabisulfite yielded no radical EPR signal above background. The intensity of Figure 3. Propofol dimer quinone in sulfite propofol emulsion and EDTA propofol emulsion during a the EPR spectrum of the sulfite anion simulated 12-hr infusion. Values represent the mean (ϩSD) of triplicate determinations from three radical spin adduct on a simulated intra- 50-mL vials of each propofol formulation. The vials used are listed in the legend to Figure 2. *Values venous infusion remained relatively con- were significantly different from corresponding 0-hr values. stant over 12 hrs (Fig. 5).

Crit Care Med 2003 Vol. 31, No. 3 789 DISCUSSION DMPO (Fig. 6) (13, 15). This radical was environments is consistent with several detected at a relatively constant level for lines of evidence. For example, sulfite has This study demonstrates that the sul- the entire 12-hr time frame of the simu- been shown to cause the peroxidation of fite propofol emulsion, but not the EDTA lated intravenous infusion. lipids in various lipid and fatty acid emul- propofol emulsion, undergoes chemical The increases of propofol dimer and sions (16–18). It has also been shown to changes during a simulated intravenous propofol dimer quinone in sulfite propo- cause the peroxidation of lipids in propo- infusion. Compounds identified as propofol emulsion demonstrate that sulfite fol emulsions (11). Mechanisms have fol oxidation products, propofol dimer from metabisulfite is in some manner been proposed that involve sulfite- and propofol dimer quinone, increased creating a strong oxidative environment facilitated formation of lipid radicals (17). during the simulation. The occurrence when air is introduced. Although sulfite These lipid radicals, generated at unsat- and identities of propofol dimer and primarily functions as an antioxidant (4), urated moieties, result in the formation propofol dimer quinone are confirmed by its ability to cause oxidation in certain of lipid peroxides. These subsequently the facts that they are extractable from propofol emulsions with an organic sol- vent, they chromatograph identically with authentic compounds by high- pressure liquid chromatography, and they have identical mass spectral and ul- traviolet-visible absorbance profiles as the standards. The sulfite free radical is confirmed by ⅐Ϫ the detection of the DMPO/SO3 radical adduct that results from the spin trapping of the sulfite radical by the spin trap

Figure 5. Relative free radical signals in sulﬁte propofol emulsion and EDTA propofol emulsion during a simulated 12-hr infusion. Values represent the means of duplicate determinations from two vials each of propofol emulsion.

Figure 4. Electron paramagnetic resonance spectra of sulﬁte propofol emulsion (a) and EDTA propofol emulsion (b) on immediate analysis after removal from intact vials. The spectra were acquired as described in MATERIALS AND Figure 6. Scheme for the formation and trapping of the sulﬁte anion radical by 5,5-dimethyl-1- METHODS. pyrroline-N-oxide(DMPO).

790 Crit Care Med 2003 Vol. 31, No. 3 ulﬁte causes oxidation of propofol S after exposure of the drug to air during a simulated intravenous infusion.

break down into more free radical or al- dehydic cleavage products (12), including Figure 7. Scheme for the formation of propofol dimer and propofol dimer quinone in propofol malondialdehyde (19), and 4-hy- emulsions. droxynonenal (20, 21). In addition, chemical studies of sulfite in the presence of oxygen have revealed emulsion, traps the sulfite radical and fol dimer quinone to convert to the col- that several sulfite-derived oxidant spe- thereby inhibits secondary radical pro- orless propofol dimer also indicates that cies form (13, 15, 22, 23). Sulfite under- cesses. emulsion yellowing is not always a reli- goes two types of oxidation, a two- The simultaneous occurrence and able indicator of the extent of propofol electron oxidation leading to the sulfate concurrent increases of both propofol oxidation. 2Ϫ 3 2Ϫ anion (SO3 SO4 ) and a one- dimer quinone and propofol dimer, indi- Even though this study evaluated a electron oxidation leading to the forma- cate that these two compounds are inter- simulated infusion from 50-mL vials over 2Ϫ tion of the reactive sulfite radical (SO3 related. Indeed, the propofol dimer qui- 12 hrs, intensive care unit patients re- 3 ⅐Ϫ SO3 ). The sulfite radical has been none, which always occurs in slightly ceiving propofol sedation at the recom- shown to react with oxygen, forming two lesser quantities than propofol dimer, is mended rates of 1–3mg·kgϪ1·hrϪ1 or less strong oxidant species, the sulfite peroxyl likely a direct oxidation product of propo- could receive propofol from a single 2Ϫ 3 Ϫ ⅐ radical (SO3 SO3OO ) and the sul- fol dimer. A one-electron oxidation of 100-mL vial over 12 hrs. For example, a 2Ϫ 3 ⅐Ϫ fate radical (SO3 SO4 ). Lesser propofol dimer will yield a propofol dimer 50-kg person receiving an average dose of quantities of superoxide are also gener- radical; however, a two-electron oxida- 1.68 mg·kgϪ1·hrϪ1 (28 ␮g·kgϪ1·minϪ1) ated (13). These species, and the sulfite tion of propofol dimer would be expected (25) will require 11.9 hrs to deplete a radical itself, may react with lipid to to yield the quinone compound (Fig. 7). 100-mL vial of 1% propofol. Longer hang cause lipid radical formation. Further- The observation that the strong reduc- times may occur at lower doses (lower more, any of these species may also di- tant dithionite causes a loss of propofol infusion rates), with smaller patients, or rectly or indirectly initiate a one-electron dimer quinone and yellow color and an with intermittent infusions from a single oxidation of propofol, yielding the propo- increase in propofol dimer confirms that vial. Of interest is that the highest rates fol radical (24) and, subsequently, propo- these compounds can interconvert, de- of increase in propofol dimer and propofol dimer and propofol dimer quinone. pending on oxidative or reducing condi- fol dimer quinone occurred from 6 to 12 Although propofol dimer and propofol tions; this observation suggests that the hrs after initiation of the simulation. Be- dimer quinone required some time to in- dimer may redox cycle. Furthermore, it cause of this result and the continued crease in quantity in this study, EPR indicates that the yellow appearance of presence of radical, these oxidation pro- analysis shows that the sulfite radical is sulfite propofol emulsions can be attrib- cesses are expected to continue and pos- rapidly formed when emulsion is with- uted to the presence of sufficient quanti- sibly accelerate after a 12-hr period. drawn from the vial. The delayed oxida- ties of propofol dimer quinone after The immediate occurrence of sulfite tion of propofol relative to sulfite radical emulsion oxidation. The violet region radical in dripped-out emulsion and its formation could be due to oxygen re- (400–430 nm) of the visible electromag- persistence in a hanging vial over 12 hrs quirements needed to form sufficient netic spectrum is complementary to yel- not only implicates sulfite radical in quantities of products, or due to physical low; therefore, a compound having an emulsion oxidation, but demonstrates factors such as phase separation of reac- absorbance band at 421–422 nm, as does that this radical is an infused chemical tants. Sulfite-derived radicals are H2O the propofol dimer quinone, will appear entity even during short propofol admin- soluble and localize in the aqueous phase yellow. Previously, it has been speculated istrations. The in vivo consequences of of the emulsion, whereas propofol and that yellowing of sulfite-containing intravenous infusion of these chemical lipid localize in the soybean oil microdro- propofol emulsion after air exposure is entities are not known. The sulfite radical plets (1). The lack of detection of other due to the presence of oxidized sulfite is a chemically reactive intermediate proposed radical species (e.g., lipid radi- species (8). However, sulfite and its oxi- (26). It can, for example, cleave albumin cal or propofol radicals) may be due to dized products are H2O soluble and will (27) and adduct to uracil (28). Sulfite the presence of large quantities of the not extract into ethyl acetate as does radical species are also thought to be sulfite radical, which would diminish any propofol dimer quinone and the yellow involved in sulfite-induced lung and neu- other radical signals, or may be due to color from a yellowed sulfite propofol ronal damage via sulfite’s ability to stim- the fact that the DMPO, when added to emulsion. The ability of the yellow propo- ulate neutrophils and activate the respi-

Crit Care Med 2003 Vol. 31, No. 3 791 ratory burst (29) and by its ability to products. Am J Health Syst Pharm 2000; antioxidant potential of propofol (2,6- potentiate peroxynitrite’s oxidant effects 57:1176–1177 diisopropylphenol). Br J Anaesth 1992; 68: (30, 31). Lipid peroxidation processes are 9. Baker MT: Yellowing of metabisulfite- 613–618 also associated with detrimental effects in containing propofol emulsion. Am J Health 25. Higgins TL, Murray M, Kett DH, et al: Trace vivo (19, 21, 32, 33). Syst Pharm 2001; 58:1042–1044 element homeostasis during continuous se- 10. Mirejovsky D: Yellowing of metabisulfite- The generic sulfite propofol emulsion dation with propofol containing EDTA versus containing propofol emulsion. Am J Health other sedatives in critically ill patients. is rated therapeutically equivalent to the Syst Pharm 2001; 58:1046–1047 Intensive Care Med 2000; 26:S422–S432 EDTA propofol emulsion (34). However, 11. Baker MT, Dehring DJ, Gregerson MS: Sul- 26. Chamulitrat W: Activation of the superoxide- we demonstrated that there are signifi- fite supported lipid peroxidation in propofol generating NADPH oxidase of intestinal lym- cant differences between the formula- emulsions. Anesthesiology 2002; 97: phocytes produces highly reactive free radi- tions in regard to chemical reactivity that 1162–1167 cals from sulfite. Free Radical Biol Med 1999; may be associated with detrimental pro- 12. Mead JF, Alfin-Slater RB, Howton DR, et al: 27:411–421 cesses. Other differences, including drug Peroxidation of fatty acids. In: Lipids, Chem- 27. Ito K, Kawanishi S: Site-specific fragmen- compatibilities, have also been observed istry, Biochemistry and Nutrition. New York, tation and modification of albumin by sul- (35). The effects of sulfite in propofol Plenum Press, 1986, pp 83–99 fite in the presence of metal ions or per- 13. Mottley C, Mason RP: Sulfate anion free rad- emulsions need to be further studied. oxidase/H2O2: Role of sulfate radical. icals formation by the peroxidation of (Bi) Biochem Biophys Res Comm 1991; 176: sulfite and its reaction with hydroxyl radical 1306–1312 CONCLUSIONS scavengers. Arch Biochem Biophys 1988; 28. Hayatsu H, Masao I: The oxygen-mediated 267:681–689 reaction between 4-thiouracil derivatives and The results presented here show that 14. Buettner GR: Spin trapping: ESR parameters bisulfite: Isolation and characterization of sulfite causes oxidation of propofol after of spin adducts. Free Radic Biol Med 1987; 1-methyluracil 4-thiosulfate as an interme- 3:259–303 exposure of the drug to air during a sim- diate in the formation of 1-methyluracil-4- 15. Hayon E, Treinin A, Wilf J: Electronic spec- sulfonate. J Am Chem Soc 1971; 93: ulated intravenous infusion. The reactiv- tra, photochemistry, and autoxidation: 2301–2306 ity of sulfite can be attributed to the for- Mechanism of the sulfite-bisulflite-pyrosul- 29. Beck-Speier I, Lenz A-G: Responses of hu- mation of a reactive sulfite radical. fite systems. The SO Ϫ,SOϪ,SOϪ, and 2 3 4 man neutrophils to sulfite. J Tox Environ SO Ϫ radicals. J Am Chem Soc 1972; 94: 5 Health 1994; 41:285–297 47–57 30. Reist M, Marshall KA, Jenner P, et al: Toxic REFERENCES 16. Southerland WM, Akogyeram CO, Toghrol F, effects of sulphite in combination with per- et al: Interaction of bisulfite with unsatur- oxynitrite on neuronal cells. J Neurochem 1. Jones C: Fundamentals of emulsions. Am J ated fatty acids. J Toxicol Environ Health 1998; 71:2431–2438 Anesthesiol 2000; 27(Suppl):12–15 1982; 10:479–491 31. Pelletier M, Savoie A, Girard D: Activation of 2. Langevin PB, Gravenstein N, Doyle TJ, et al: 17. Kaplan D, McJilton C, Luchtel D: Bisulfite Growth of Staphylococcus aureus in induced lipid oxidation. Arch Environ Health human neutrophils by the air pollutant so- diprivan and intralipid: Implications on the 1975; 30:507–509 dium sulfite (Na2SO3): Comparison with im- pathogenesis of infections. Anesthesiology 18. Lizada MCC, Yang SF: Sulfite-induced lipid mature promyelocytic HL-60 and DMSO- 1999; 91:1394–1400 peroxidation. Lipids 1981; 16:189–194 differentiated HL-60 cell reveals that Na2SO3 3. Simon RA: Food and drug additives. Immun 19. Tuma DJ: Role of malondialdehyde-acetalde- is a neutrophil but not a HL-60 cell agonist. Allergy Clin 1995; 15:489–527 hyde adducts in liver injury. Free Radic Biol Clin Immunol 2000; 96:131–139 4. Usseglio-Tomasset L: Properties and use of Med 2002; 32:303–308 32. Webster NR, Goode HF: Free radicals and sulphur dioxide. Food Addit Contam 1992; 20. Poli G, Schaur RJ: 4-Hydroxynonenal in the antioxidants in sepsis. Crit Care Med 1993; 9:339–404 pathomechanisms of oxidative stress. IUBMB 21:1770–1776 5. Kraniak JM, Shelef LA: Effect of ethylenedia- Life 2000; 50:315–321 33. Helbock HJ, Motchnik PA, Ames BN: Toxic minetetraacetic acid (EDTA) and metal ions 21. Koh YH, Yoon SJ, Park JW: Lipid peroxida- hydroperoxides in intravenous lipid emul- on growth of Staphylococcus aureus 196E in tion product-mediated DNA damage and mu- sions used in preterm infants. Pediatrics culture media. J Food Sci 1988; 53:910–913 tagenicity. J Biochem Mol Biol 1997; 30: 1993; 91:83–87 6. Baker MT: Comparison of emulsion chemis- 188–193 34. Center for Drug Evaluation and Research, try between sulfite-containing propofol and 22. Sun X, Shi X, Dalal NS: Xanthine oxidase/ Food and Drug Administration, US Depart- propofol with EDTA. Am J Anesthesiol 2000; hydrogen peroxide generates sulfur trioxide ment of Health and Human Services: Elec- ⅐Ϫ Ϫ2 27(Suppl):19–21 anion radical (SO3 ) from sulfite (SO3 ). tronic Orange Book: Approved Drug Prod- 7. Redhead HM, Jones CB, Bobey DG: Pharma- FEBS Lett 1992; 303:213–216 ucts with Therapeutic Equivalence ceutical and antimicrobial differences be- 23. Lavoie JC, Lachance C, Chessex P: Antiper- Evaluations (May 2002). Available at: http:// tween propofol products. Am J Health Syst oxide activity of sodium metabisulfite: A dou- www/fda.cder/ob/default.htm Pharm 2000; 57:1174–1176 ble-edged sword. Biochem Pharmacol 1994; 35. Trissel LA: Drug compatibility differences 8. Mirejovsky D, Ghosh M: Pharmaceutical and 47:871–876 with propofol injectable emulsion products. antimicrobial differences between propofol 24. Murphy PG, Myers DDS, Davies MJ, et al: The Crit Care Med 2001; 29:466–467

792 Crit Care Med 2003 Vol. 31, No. 3