Secondary Logo

Journal Logo

Anesthetic Pharmacology: Preclinical Pharmacology: Clinical Pharmacology: Research Reports

Sevoflurane Formulation Water Content Influences Degradation by Lewis Acids in Vaporizers

Kharasch, Evan D. MD, PhD*; Subbarao, Gowdahalli N. PhD; Cromack, Keith R. PhD; Stephens, Dennis A. PhD§; Saltarelli, Mario D. MD, PhD

Editor(s): Durieux, Marcel E.; Gin, Tony

Author Information
doi: 10.1213/ane.0b013e3181a3d72b
  • Free


Sevoflurane was first introduced in 1990 by Maruishi Pharmaceuticals in Japan, and subsequently marketed in 1995 by Abbott Laboratories in the United States (US) as Ultane® and worldwide as Sevorane®. Generic versions of sevoflurane were introduced nearly a decade later (2006) and are considered by regulatory agencies to be therapeutically equivalent to proprietary sevoflurane. The currently marketed formulations of sevoflurane differ with respect to their methods of synthesis, impurities, the containers in which they are sold, and their formulation, specifically, their water content.1 The manufacturing impurities are thought not to be quantitatively and clinically significant, as long as they remain low.1

Sevoflurane, like all currently used volatile anesthetics, is susceptible to chemical degradation. One type of degradation is caused by Lewis acids (such as metal halides) and results in potentially toxic compounds such as hydrogen fluoride. This corrosive acid, if inhaled, may cause respiratory irritation or pulmonary hemorrhage.2,3 In the US, the Occupational Safety and Health Administration limits hydrofluoric acid exposure to 3 ppm (over an 8 h average). Hydrofluoric acid is also highly reactive and corrosive.2,3

One of the first incidences of Lewis acid-mediated sevoflurane degradation occurred in 1996, in which several bottles were found to contain cloudy drug with a pungent odor, marked acidity (pH < 1) and high levels of fluoride (863 ppm).4,5 Although no patients were exposed to the degraded drug, it prompted a thorough evaluation of causality, and methods to prevent such degradation. Investigators at Abbott determined that increasing the water content in sevoflurane formulations decreased Lewis acid-dependent sevoflurane degradation,6 and as a result, the formulation of Abbott sevoflurane was modified to contain at least 300 ppm water. This amount was sufficient to inhibit Lewis acid-mediated degradation and the formation of toxic degradants under normal clinical conditions. Currently marketed sevoflurane formulations contain <130 ppm water. It is unknown whether these generic sevoflurane formulations are susceptible to Lewis acid-mediated degradation in clinical use situations.

Commonly used anesthesia vaporizers have metal surfaces that may contain potential Lewis acids (metal oxides). Recent events suggest that vaporizers containing such Lewis acids may degrade sevoflurane. European Medicines and Healthcare Products Regulatory Agency notices* and reports in abstract form,7,8 stated that some Penlon Sigma Delta sevoflurane vaporizers (distributed by Baxter) were found to interact with low-water sevoflurane formulations, resulting in etching of the vaporizer sight glass and corrosion of the metal filling port shoe. Etching of the sight glass resulted in significant interference with the ability to read the sevoflurane liquid levels in the vaporizer. In both European Agency notices, it was recommended that the vaporizers be removed from use. Sight glass etching suggests the potential formation of hydrofluoric acid. The degradation products in the vaporizers in the above reports were not, however, reported.

Although the currently marketed lower-water and higher-water sevoflurane formulations are considered pharmaceutically equivalent by the US Food and Drug Administration, their water content differs. The purpose of this investigation was to assess the influence of sevoflurane formulation (water content) and type of vaporizer on the stability and degradation of sevoflurane. Three commonly used vaporizers were assessed. Sevoflurane degradation was determined by measuring total impurities and hydrofluoric acid formation in three sevoflurane formulations stored in these vaporizers for various periods of time.


All experiments and laboratory analyses were performed at PPD Development, LP (Middleton, WI). Three commercially available sevoflurane formulations, containing 19 ppm water (Eraldin, Laboratorios Richmond/Minrad, Argentina), 57 ppm water (generic sevoflurane, Baxter, US), and 352 ppm water (Ultane, Abbott), and three commercially available vaporizers (GE/Datex-Ohmeda Tec 7, GE: Vapor 2000, Draeger; and Sigma Delta, Penlon) were evaluated. Pairs of new vaporizers were filled with one 250 mL newly opened container of each sevoflurane product using a clean and dry filling adapter. Filling and sampling were performed in a low-humidity glove box to minimize exposure to ambient humidity, and the vaporizers were sealed. Experiments were conducted under accelerated stability conditions (40°C, 15% relative humidity) that model longer-term room temperature storage conditions.

Samples of sevoflurane for analyses (40 mL) were drained into polyethylene napthalate bottles for analysis immediately after filling and after 1, 2, and 3 wk. Water content was determined by the Karl Fischer method. Liquid-liquid extraction of sevoflurane (10 mL) was performed using an equal volume of water. The fluoride concentration of the aqueous layer was determined using a fluoride ion-specific electrode (Orion 96–09), with a lower limit of quantification of 0.4 ppm. pH was determined potentiometrically at ambient conditions. Samples were analyzed by gas chromatography-flame ionization detection for quantification of sevoflurane and degradation products. The gas chromatography system (Model 6890, Agilent, Palo Alto, CA) used a Quadrex column (30 m × 320 μm, 3 μm), helium carrier gas (1.0 mL/min), 200°C injector, 225°C detector, and an oven temperature of 40°C for 10 min, increased at 10°C/min to 200°C and held for 14 min. Quantitation was performed using an external standard solution of sevoflurane. Individual impurities were quantified by peak integration assuming a similar detector response to that of sevoflurane, using the equation:

Total degradant formation was the sum of all peaks (excluding sevoflurane).

An additional experiment was conducted to determine the water contents of two different sevoflurane formulations (Abbott and Minrad) in two types of vaporizers (GE Tec 7 and Penlon Sigma Delta), to address whether and how water contents in the vaporizers could change over time. New vaporizers were used, and all procedures performed in a dry box, as before. Each type of vaporizer was filled with each type of sevoflurane, a small (15 mL) sample was drained for water analysis (Karl Fischer), the vaporizer kept in a dry box overnight, another small sample drained for water analysis, and the sevoflurane was then vaporized using dry nitrogen (10 L/min), 4% setting, until the liquid level was reduced down to about 2–3 mm (Tec 7) or 4–5 mm (Penlon) above the minimum mark in the site window (representing 10%–15% above the minimum mark for the vaporizer, to ensure sufficient sample amount for water analysis), after which another sample was drained for water analysis. The vaporizer was refilled from a newly opened bottle of sevoflurane. This process was repeated daily for 2 wk, with the goal of achieving equilibrium with respect to water content. This was verified by constant water concentrations in each vaporizer for the last few says of the second week. With no further vaporization or refilling, a small sample was drained for water analysis each week for 4 wk. At week 4, the vaporizer was fully drained. Any residual water in the vaporizer was then quantified by filling the vaporizer to the maximum mark with dry methanol (of which the water content was determined before the filling and subtracted from the final water content) and kept overnight. The next morning, vaporizers were tilted so that the methanol would contact as much of the internal vaporizer surface as possible. The methanol was then drained, and water content determined.


Sevoflurane degradation varied substantially between vaporizers and between lower- versus higher-water content formulations. The Penlon Sigma Delta was the only vaporizer associated with sevoflurane degradation. Substantial amounts of fluoride were found in both lower-water sevoflurane formulations after 2 wk in the Penlon Sigma Delta vaporizer, increasing further to 600 ppm after 3 wk (Fig. 1A). In contrast, the higher-water sevoflurane formulation in the Penlon Sigma Delta contained negligible amounts of fluoride at all time points. None of the three sevoflurane formulations stored in the Drager Vapor 2000 (Fig. 1B) or the GE/Datex-Ohmeda Tec (Fig. 1C) vaporizers were found to contain any measurable fluoride at any time point.

Figure 1.:
Influence of sevoflurane water content on degradation to fluoride in anesthetic vaporizers. Various sevoflurane formulations were stored at 40°C in (A) Penlon Sigma Delta vaporizers, (B) Draeger Vapor 2000, and (C) GE/Datex-Ohmeda Tec 7 vaporizers, and aliquots removed weekly for analysis of fluoride content. Fluoride concentrations are shown for each of two vaporizers. Each data point is a single determination. Fluoride concentrations in Draeger Vapor 2000 and the GE/Datex-Ohmeda Tec7 vaporizers were non-detectable or below the limit of quantification (<0.4 ppm).

Substantial decreases in pH of the lower-water sevoflurane formulations were observed after 2 and 3 wk storage in the Penlon Sigma Delta vaporizers (Table 1). In contrast, only small decreases in pH were observed during storage of the higher-water sevoflurane formulation in the Sigma Delta vaporizers. The pH of all sevoflurane formulations remained stable during storage in the Draeger Vapor 2000 and the GE/Datex-Ohmeda Tec 7 vaporizers (Table 1).

Table 1:
Influence of Sevoflurane Water Content on Degradation in Vaporizers

Gas chromatographic analysis was used to determine total degradant formation in addition to fluoride. The lower-water sevoflurane formulations in Penlon Sigma Delta vaporizers exhibited a time-dependent increase in average total degradants (Fig. 2A), which exceeded 68,000 ppm after 3 wk, compared with <1400 ppm in the higher-water sevoflurane. There was no change in total degradant concentration with any sevoflurane formulation (lower- or higher-water content) stored in the Draeger Vapor 2000 or the GE/Datex-Ohmeda Tec 7 vaporizers (Figs. 2B and C). Of note, total degradant concentration at baseline in the lower-water sevoflurane (Laboratorios Richmond/Minrad, Argentina) was approximately 400 ppm, whereas that in the lower-water sevoflurane by US Baxter and the higher-water sevoflurane by Abbott was <60 ppm.

Figure 2.:
Influence of sevoflurane water content on formation of total degradants after storage in anesthetic vaporizers. Experimental conditions and vaporizers are the same as described in the legend to Figure 1. Results are the average for both vaporizers.

Visual inspections of the Penlon Sigma Delta vaporizers revealed substantial etching of the sight glass and metal filler shoe observed after 3 wk of storage of the lower-water sevoflurane formulations (Fig. 3). No changes were observed in the vaporizers containing higher-water sevoflurane. Corrosion was not quantified, nor was corrosion from various low-water sevoflurane formulations compared.

Figure 3.:
Penlon vaporizer destruction resulting from degradation of lower-water sevoflurane. A, Initial vaporizer sight glass. B, No change in the sight glass following storage of higher-water sevoflurane (Ultane, Abbott) for 3 weeks. C and D, The Penlon vaporizer sight glass showed substantial etching following 3 weeks of storage of the lower-water sevoflurane formulations by two manufacturers.

Analyses of water content in the sevoflurane sampled from vaporizers revealed time-dependent changes in some formulations and vaporizers. In general, the water content of lower-water sevoflurane formulations decreased over time in the Penlon Sigma Delta vaporizers, while in contrast it increased over time in the Draeger Vapor 2000 and GE/Datex-Ohmeda Tec 7 vaporizers (Table 1). In general, the water content of higher-water sevoflurane remained above 150 ppm in all vaporizers.

An additional experiment was conducted to determine the water contents of two different sevoflurane formulations in two types of vaporizers, to address the question of why water contents in the vaporizers changed over time, and particularly, how sevoflurane water content could increase over the experimental period (and what was the source of the water) (Table 2). Daily vaporization and refilling achieved constant water content in each vaporizer and sevoflurane combination by day 15. Without further vaporization and refilling, sevoflurane water content was relatively unchanged (Tec 7/Abbott Ultane) or increased slowly (Tec 7/Minrad Eraldin and both Penlon vaporizers) over the next month. Fully draining the vaporizer after 1 month recovered more water. Rinsing the vaporizer with methanol to extract any remaining water recovered 3–30 times more water.

Table 2:
Water Content of Sevoflurane and Vaporizers


These findings demonstrate that water content in the various sevoflurane formulations profoundly influenced sevoflurane degradation in Penlon Sigma Delta vaporizers. Formation of hydrofluoric acid (measured as decreased pH and increased fluoride concentration), as well as total degradant concentrations, was substantially higher in the lower-water formulations as compared with the higher-water formulation over the course of the 3 wk observational study. Moreover, lower-water sevoflurane degradation was accompanied by physical damage to the vaporizers, such as corrosion of the filler and etching of the sight glass. In addition to the formation of hydrogen fluoride, a preliminary experiment showed that low-water sevoflurane degradation also resulted in the formation of hexafluoroisopropanol and various other Lewis acid degradants (results not shown). Thus lower-water sevoflurane degradation was consistent with a Lewis acid-mediated process, based on the identity of the degradation products. In contrast, degradation of higher-water sevoflurane was negligible. No meaningful production of hydrofluoric acid or other degradants occurred when the sevoflurane water content exceeded 150 ppm within the vaporizer. These results demonstrate that sufficient amounts of water prevent Lewis acid-mediated sevoflurane degradation.

Lewis acid-mediated degradation was not observed with either the lower- or the higher-water sevoflurane formulations stored in GE/Datex-Ohmeda Tec 7 or Draeger Vapor 2000 vaporizers. It is unclear why the Penlon Sigma Delta vaporizers were associated with much greater sevoflurane degradation than the other vaporizers. This may be the result of differing amounts of Lewis acid containing surfaces in the respective vaporizers. The etching and corrosion observed in the present study is similar to that of recent reports of clinical vaporizer damage7,8 that culminated in a recall of these vaporizers.§ The similarities to the present observations suggest that the clinical Penlon vaporizer damage likely resulted from Lewis acids in the Penlon Sigma Delta vaporizers interacting with sevoflurane in the absence of sufficient amounts of protective water, resulting in hydrogen fluoride formation.

An interesting question was the cause of the varying water content in sevoflurane over time (Table 1). Results of a follow-up experiment suggest that the increase in sevoflurane water content in sealed vaporizers originates from the vaporizers. Potential sources include the “soft parts”, such as a cotton wick, which is relatively hydrophilic, gaskets, or other materials. Residual moisture in the vaporizers after manufacture could readily equilibrate over time with sevoflurane, accounting for the increased sevoflurane water content. Not all vaporizer water equilibrates with sevoflurane, as more was extractable by methanol (which is miscible with water whereas sevoflurane is not). Therefore the change in vaporizer water content in the first experiments is due to absorption or desorption of water trapped within vaporizer components which is in fluid contact with sevoflurane, and that an equilibrium is reached between water in the components and sevoflurane, as the drug sits in the vaporizer.

Alternative hypotheses for the source of water have no support. For example, there is no known reaction of sevoflurane that should liberate water. Moreover, it is unlikely that the water comes from the atmosphere, otherwise, one would also expect the sevoflurane to evaporate through whatever portal permitted entry of water vapor from room air. In addition, it is highly unlikely that some (as yet unidentified) potential impurities in the generic formulations might liberate water (by some as yet unknown reaction). This is because generic sevoflurane(s) (while having different impurity profiles due to their various manufacturing schemes) have <100 ppm of any single impurity, which (assuming some reaction did exist) would produce water only 10’s of ppm (w/w) due to stoichiometric considerations. In fact, much higher water concentrations were measured.

The data also show that the Penlon vaporizer contains substantially less total water than the GE vaporizer. This may also contribute to the greater degradation of low water sevoflurane in the Penlon vaporizer. In addition, higher water sevoflurane can actually increase the vaporizer water concentration compared with lower water sevoflurane. This may also contribute to the lesser degradation of higher versus lower water sevoflurane in the Penlon vaporizer.

These results should be considered preliminary for several reasons and hence interpreted with caution. First, the observational analyses were performed only in duplicate and only for up to 3 wk. In many practice settings sevoflurane may sit in vaporizers for even longer periods of time with an even greater potential for degradation to occur with lower-water formulations. Second, the experiment was conducted under accelerated conditions thought to mimic longer storage times (elevated temperature), in compliance with international guidelines on stability and degradation. The degradation profiles of sevoflurane after longer periods of residence in vaporizers, and at room temperature, is unknown. It is also unknown whether elevated temperatures in pediatric operating rooms, such as those in which sevoflurane degradation was first reported, influences stability in vaporizers.7 Lastly, any changes in vaporizer manufacturing or composition may impact degradation.

The potential patient safety implications of these observational study results ethically mandated their reporting before, rather than after, replicating the experiments with larger numbers and more extensive degradant analysis. Such analyses are in progress. In the meantime, practitioners should be aware of the fact that the lower-water and higher-water formulations of sevoflurane, while considered therapeutically equivalent, may not have the same patient safety profile. It is imperative to emphasize that the greater stability of the higher-water sevoflurane formulation should not be misinterpreted to imply that practitioners can or should add water to a lower-water sevoflurane formulation in an attempt to prevent potential degradation. The effects of such a process are unknown and untested, and water should only be added under proper controlled manufacturing conditions.

In summary, sevoflurane formulations containing low amounts of water can interact with metal oxides (Lewis acids) in certain commercial vaporizers, resulting in substantial amounts of degradation to toxins such as hydrofluoric acid. This did not occur during storage of higher-water sevoflurane formulations. The differences in water content of sevoflurane formulations and associated degradation present a potential patient safety issue.


1. Baker M. Sevoflurane: are there differences in products? Anesth Analg 2007;104:1447–51
2. Dalbey W, Dunn B, Bannister R, Daughtry W, Kirwin C, Reitman F, Steiner A, Bruce J. Acute effects of 10-minute exposure to hydrogen fluoride in rats and derivation of a short-term exposure limit for humans. Regul Toxicol Pharmacol 1998;27:207–16
3. Bertolini J. Hydrofluoric acid: a review of toxicity. J Emerg Med 1992;10:163–8
4. Leary JP. Sevoflurane containing toxic acid prompts recall. Contaminated sevoflurane use reported from NY State [Letter to the Editor]. APSF Newsl 1996;11:37–9
5. Callan CM. Sevoflurane containing toxic acid prompts recall. Sevo manufacturer outlines circumstances [Response]. APSF Newsl 1996;11:37–9
6. McLeskey CH. Anesthesiologist executive reports how Abbott made sevoflurane safer: water stops formation of highly toxic acid. APSF Newsl 2000;15:39
7. O’Neill B, Hafiz MA, DeBeer DA. Corrosion of Penlon sevoflurane vaporizers. Anaesthesia 2007;62:421
8. Gupta A, Ely J. Faulty sevoflurane vaporizer. Anaesthesia 2007;62: 421

* Accessed May 19, 2007; and Accessed May 19, 2007.
Cited Here

†U.S. Department of Health and Human Services: Guidance for Industry: Q1A (R2) Stability Testing of New Drug Substances and Products. November 2003, Revision 2. Accessed May 7, 2007. and International Conference on Harmonization (ICH) Q1A(R2) Stability Testing Guidelines: Stability Testing of New Drug Substances and Products. The European Agency for the Evaluation of Medicinal Products. CPMP/ICH/2736/99. London, 20 February 2003.
Cited Here

‡In this preliminary experiment, sevoflurane samples were removed from the vaporizers and, using gas chromatography, compared with samples that had been exposed to alumina (a Lewis acid). The chromatograms of the lower water sevoflurane products that had been exposed to alumina contained numerous peaks representative of various degradants, with these peaks not being present in the chromatograms of “fresh” drug. Two different formulations of lower-water sevoflurane removed from Penlon Sigma Delta vaporizers after 3 weeks also contained numerous degradant peaks, and these chromatograms were qualitatively similar to those from Lewis acid (alumina)-degraded lower-water sevoflurane. In contrast, higher-water sevoflurane removed from Penlon Sigma Delta vaporizers after 3 weeks showed few additional peaks compared with fresh drug. Hexafluoroisopropanol was identified as one of the lower water sevoflurane degradants common to both alumina and Penlon Sigma Delta vaporizers, based on retention time.
Cited Here

§Medicines and Healthcare Products Regulatory Agency: Vaporizers: vaporizer—Penlon—Sigma Delta Vaporiser—Sevoflurane [updated: Document No:2006/008/010/401/007], 04 December 2006, pp
Cited Here

‖According to Penlon, it has since redesigned the Sigma Delta vaporizer, using plastic coated internal surfaces to remove potential Lewis acid production, and eliminate the issues with potential of degradation (Anesthesia Patient Safety Foundation Newsletter, 22:84, Winter 2007–2008). We have not evaluated the stability of sevoflurane formulations with this redesigned vaporizer.
Cited Here

© 2009 International Anesthesia Research Society