Secondary Logo

Journal Logo

Occupational Chronic Sevoflurane Exposure in the Everyday Reality of the Anesthesia Workplace

Herzog-Niescery, Jennifer MD*; Botteck, Nikolaj Matthias MD*; Vogelsang, Heike MD*; Gude, Philipp MD*; Bartz, Horst DrEng; Weber, Thomas Peter MD*; Seipp, Hans-Martin MD, DiplEng

doi: 10.1213/ANE.0000000000001015
Patient Safety: Research Report

BACKGROUND: Although sevoflurane is one of the most commonly used volatile anesthetics in clinical practice, anesthesiologists are hardly aware of their individual occupational chronic sevoflurane exposure. Therefore, we studied sevoflurane concentrations in the anesthesiologists’ breathing zones, depending on the kind of induction for general anesthesia, the used airway device, and the type of airflow system in the operating room. Furthermore, sevoflurane baselines and typical peaks during general anesthesia were determined.

METHODS: Measurements were performed with the LumaSense Photoacoustic Gas Monitor. As we detected the gas monitor’s cross-sensitivity reactions between sevoflurane and disinfectants, regression lines for customarily used disinfectants during surgery (Cutasept®, Octeniderm®) and their alcoholic components were initially analyzed. Hospital sevoflurane concentrations were thereafter measured during elective surgery in 119 patients. The amount of inhaled sevoflurane by anesthesiologists was estimated according to mVA = cVA × V × t × ρVA aer.

RESULTS: Induction of general anesthesia stopped after tracheal intubation with the patient’s expiratory sevoflurane concentration of 1.5%. Thereby, inhalational inductions (INH) caused higher sevoflurane concentrations than IV inductions (mean [SD]:

[ppm] INH 2.43 ±1.91 versus IV 0.62 ± 0.33, P < 0.001; mVA [mg] INH 1.95 ± 1.54 versus IV 0.30 ± 0.22, P < 0.001). The use of laryngeal mask airway (LMA™) led to generally higher sevoflurane concentrations in the anesthesiologists’ breathing zones than tracheal tubes (

[ppm] tube 0.37 ± 0.16 versus LMA™ 0.79 ± 0.53, P = 0.009;

[ppm] tube 1.91 ± 0.91 versus LMA™ 2.91 ± 1.81, P = 0.057; mVA [mg] tube 1.47 ± 0.64 versus LMA™ 2.73 ± 1.81, P = 0.019). Sevoflurane concentrations were trended higher during surgery in operating rooms with turbulent flow (TF) air-conditioning systems compared with laminar flow (LF) air-conditioning systems (

[ppm] TF 0.29 ± 0.12 versus LF 0.13 ± 0.06, P = 0.012; mVA [mg/h] TF 1.16 ± 0.50 versus LF 0.51 ± 0.25, P = 0.007).

CONCLUSIONS: Anesthesiologists are chronically exposed to trace concentrations of sevoflurane during work. Inhalational inductions, LMA™, and TF air-conditioning systems in particular are associated with higher sevoflurane exposure. However, the amount of inhaled sevoflurane per day was lower than expected, perhaps because concentrations in previous measurements could be overestimated (10%–15%) because of the cross-sensitivity reaction.

From the *Department of Anesthesiology, Katholisches Klinikum Bochum, St. Josef- and St. Elisabeth Hospital, Ruhr-University of Bochum, Bochum, Germany; and Department of Hygiene, Environmental Engineering and Biotechnology, University of Applied Sciences, Giessen, Germany.

Accepted for publication August 6, 2015.

Funding: None.

The authors declare no conflicts of interest.

LMA is a registered trade mark of The Laryngeal Mask Company Limited, an affiliate of Teleflex Incorporated.

Reprints will not be available from the authors.

Address correspondence Jennifer Herzog-Niescery, MD, Department of Anesthesiology, St. Josef-Hospital Bochum, Gudrunstraße 56, D-44791 Bochum, Germany. Address e-mail to

Although sevoflurane is one of the most commonly used volatile anesthetics in clinical practice, anesthesiologists are hardly aware of their individual occupational chronic sevoflurane exposure. This is particularly surprising, as neither the harmfulness nor harmlessness of low-concentration sevoflurane exposure in anesthetic personnel has been convincingly proven.

Most studies have shown no evidence of genetic toxicity after sevoflurane exposure in structural chromosome aberration assays or in the Ames Salmonella mutagenesis test.1–3 However, increased formation of sister chromatid exchanges were detected in animal studies, as well as in anesthetic personnel after chronic sevoflurane exposure.2,4 Furthermore, authors discussed congenital anomalies in children, whose mothers were exposed to sevoflurane.5 Consequently, this debate remains controversial.

To date, a worldwide uniform time-weighted average exposure threshold limit concentration for sevoflurane has not been defined. In the United States, the National Institute for Occupational Safety and Health recommends an exposure limit of 2.00 ppm for sevoflurane for the duration of a procedure (60 minutes ceiling limit that should not be exceeded during any part of the workday).6,7 Some (European) countries propose an 8-hour time-weighted average exposure limit of 7.00 ppm for sevoflurane, whereas others do not recommend any “acceptable” concentration for waste sevoflurane in the air of operating rooms (ORs).8 Nevertheless, occupational chronic sevoflurane exposure should obviously be as low as possible. Therefore, it is essential to determine sevoflurane exposure during anesthesia.

Few studies have focused on sevoflurane concentration in the air during inhalational inductions, or while different airway devices were used, and in those that have, most used a photoacoustic gas monitor to determine sevoflurane concentrations.9–13 Photoacoustic gas monitoring has been used for years in the field of anesthesia.14 This method selectively detects trace concentrations of anesthetic gas and enables highly accurate, reliable, and stable real-time measurements without sample preparation.15 However, in 1993, a small sensitivity to alcohol was detected for infrared analyzers when using isoflurane but not for photoacoustic analyzers.15 Cross-sensitivity reaction for photoacoustic analyzer to alcohol with a filter for sevoflurane had not been tested.

Different alcohols are main components of disinfectants, which are frequently used in the OR. Therefore, we analyzed, in a preclinical experimental setting, potential cross-sensitivity reactions between the gas monitor (filter for sevoflurane) and isopropyl alcohol, ethyl alcohol, N-propanol, Cutasept® (isopropyl alcohol 72 vol%, BODE Chemie, Hamburg, Germany) and Octeniderm® (N-propanol 30 vol%; isopropyl alcohol 45 vol%, Schülke & Mayr GmbH, Norderstedt, Germany). Measurements revealed a distinctive sensitivity to alcohols and disinfectants, so that measured sevoflurane concentrations of previously conducted studies should be re-evaluated.

Therefore, we first analyzed the alcohols’ regression lines in an experimental setting, so that sevoflurane concentrations in hospital can be calculated, even if disinfectants were used during measurements (considering cross-sensitivity).

Second, we studied sevoflurane concentrations in the anesthesiologists’ breathing zones from induction of anesthesia through surgery until the patient reached the recovery room (described as “sevoflurane baseline”), to get an impression of sevoflurane exposure during routine work.

Third, we focused on situations associated with higher sevoflurane concentrations (inhalational induction, laryngeal mask). Here, measurements from previously published studies were re-evaluated.

Fourth, 2 different airflow technologies in the OR were compared (supply air systems with laminar flow [LF] versus turbulent mixed flow).

Last, the amount of sevoflurane inhaled by the anesthesiologist was estimated, depending on sevoflurane concentration, density of sevoflurane at 20°C, respiratory minute volume, and exposure time.

Back to Top | Article Outline


The study was approved by the appropriate IRB in 2012 (4494-12, Bochum, Germany). The requirement for written informed consent was waived by the IRB.

Back to Top | Article Outline

Method of Gas Measurement

All tests were performed with the LumaSense Photoacoustic Gas Monitor Innova 1412 (Innova AirTech Instruments, LumaSense) which selectively detects sevoflurane in the air by using the photoacoustic infrared detection method. This method involves aspirating air into a measurement chamber, where light from a pulsed infrared laser passes through a narrow-band optical filter and irradiates gas molecules. The resulting volume expansion, as well as the volume contraction of the aspirated air, leads to pressure variations within the measurement chamber, which can be detected as acoustic waves by microphones.

In this study, we used the optical filter UA 0972 for sevoflurane with a center wavelength of 8.8 μm and a bandwidth of 6.0%. The calibration with a dry zero level gas resulted in an average value of 6.83E-03 ppm ± 5.00E-03 ppm. The detection limit is the minimal sevoflurane concentration that produces an observable response, which is twice the noise signal of the sevoflurane concentration when measured in dry air (0.01 ppm). The upper detection limit is 100.00 ppm. For all measurements, a sample integration time of 5 seconds was used. The interval of measurement depended on the sample integration time, the size of the measurement chamber, and the length of the gas sampling line (8 m). Here, the flushing of the measurement chamber and the gas sample line lasted 30 seconds, resulting in measurement intervals of 35 seconds. According to the manufacturer, the reproducibility of the measured sevoflurane concentration by the gas monitor is ±1% of the measured value.

Back to Top | Article Outline

Cross-Sensitivity Reaction

Tests to verify the photoacoustic analyzer’s sensitivity to alcohols and disinfectants were performed at the University of Applied Sciences, Giessen, Germany.

A chafing dish was heated to 37°C, and a metal bar was placed on top. Then, 25 mL of different alcohols (isopropyl alcohol, ethyl alcohol, N-propanol) or 25 mL disinfectant (Cutasept, Octeniderm), which combine a mixture of the previously tested alcohols in different concentrations, was distributed on the metal bar. A gas sample line was fixed 30 cm above the plate and connected to the gas monitor with a selectively calibrated filter for sevoflurane. Five measurements were performed with each agent. Thus, the maximal “sevoflurane concentration,” as reported by the gas monitor, was documented, as well as decay curves down to 10% of the maximal concentration as a function of time. In addition, dose-dependent decay curves for isopropyl alcohol were analyzed (25, 30, 35, 40, 45, and 50 mL; 5× each).

Back to Top | Article Outline

Measurements in Hospital

The clinical trial was conducted in 2 German University Hospitals during elective surgery in 119 otherwise healthy adults (ASA physical status I–II) aged 18 to 60 years undergoing tonsillectomy (n = 40), knee arthroscopy (n = 55), or neck surgery (n = 24). None of them was obese or had a history of lung disease.

Back to Top | Article Outline

Induction of Anesthesia

The anesthesiologists’ sevoflurane exposure during IV versus inhalational induction of anesthesia was analyzed. A gas sample line was fixed at the shoulder of the anesthesiologist, so that the end of the sample line was within a 10-inch (25-cm) radius of the anesthesiologist’s nose and mouth, representing the individual breathing zone. All inductions were performed in an OR with LF.

For IV inductions, patients received 100% oxygen (fresh gas flow of 8 L/min) via facial mask until an expiratory O2 fraction of 0.8 had been reached. Subsequently, sufentanil (Sufenta mite®; 0.5–2 μg/kg Janssen-Cilag GmbH, Neuss, Germany), propofol (Propofol Fresenius®, 1%; 1.5–2.5 mg/kg Propofol Fresenius 1%, Fresenius Kabi Austria, Graz, Austria), and atracurium (Tracrium®; 0.3–0.6 mg/kg Hameln Pharmaceuticals, Hameln, Germany) were administered, always followed by tracheal intubation. Tubes were connected to the ventilator and patients obtained sevoflurane (Sevorane®, AbbVie GmbH, Germany) in 100% oxygen (flow 2 L/min). Measurements stopped when the patients’ expiratory sevoflurane concentration reached 1.5%.

Inhalational inductions were performed with 8.0% sevoflurane (inspiratory concentration) in 100% oxygen (flow 10 L/min) by facial mask until lid-lash reflexes were lost. Noticeable gas leakages of the mask (smell of sevoflurane) were documented. Venous catheters were placed, and sevoflurane concentrations were reduced to 4.0% (inspiratory concentration). Sufentanil, propofol (1, and atracurium were administered. During airway manipulation, masks were removed from the patient’s face and the fresh gas flow reduced to 0.2 L/min. Tracheal intubation succeeded in all patients at the first attempt. Tubes were connected to the respirator and patients obtained sevoflurane (1.5% end-tidal) in 100% oxygen (flow 2 L/min).

Back to Top | Article Outline

Airway Devices

Sevoflurane concentrations in the air of turbulent mixed flow OR were studied depending on the chosen airway device (tube versus laryngeal mask [LMA™]), with a gas sample line fixed within the anesthesiologists’ breathing zones (n = 45) or at a distance of 30 cm above the patient’s head (n = 10).

Figure 1

Figure 1

IV inductions took place in a separated induction room. Patients were randomized to receive either a tube (internal diameter 7.0–8.0 mm) or an LMA™ (sized 4–5) made from polyvinyl chloride (Fig. 1). Cuffs were inflated with the recommended volume of air (tube approximately 5 mL; LMA™ 30–40 mL), resulting in cuff pressures between 20 to 30 cm H2O (tube) and 35 to 50 cm H2O (LMA™). Cuff pressures were kept within this range by using a VBM manometer (VBM Medizintechnik GmbH, Sulz a. N., Germany). Ventilation was performed in volume-controlled ventilator mode with a tidal volume of 6 to 8 mL/kg ideal body weight and a maximal pressure of 20 cm H2O.

Back to Top | Article Outline

Sevoflurane Baseline

The anesthesiologists’ sevoflurane exposure was analyzed during routine narcoses in 22 patients undergoing knee arthroscopy from induction of anesthesia through surgery, until the patient reached the recovery room. The gas sample line was fixed within the individual breathing zone. Measurements were conducted every 35 seconds. Patients received an IV induction in a separated induction room followed by tracheal intubation. Maintenance of anesthesia in the OR was performed with sufentanil and sevoflurane (1.5%–2.5% end-tidal). After extubation in the OR, the patient was brought to the recovery room.

Back to Top | Article Outline

Airflow System

Supply air systems with turbulent mixed flow or unidirectional LF have been 2 of the most commonly used airflow systems in ORs for approximately 50 years.16 Here, we compared the anesthesiologists’ and the surgeons’ sevoflurane exposure during balanced anesthesia, depending on the used airflow system in the OR (Fig. 2). All patients obtained cuffed tracheal tubes. Measurements started with the skin incision and stopped with the wound closure.

Figure 2

Figure 2

ORs at hospital I were equipped with turbulent mixed flow air-conditioning systems. Clean supply air enters the OR through 2 small circular air outlets from the ceiling with a high primary impulse (supporting jets), flows diagonally toward the patient, and is mixed on this way with clean air entering from the surrounding perforated steel plates. Because of the turbulences caused by this mixing, the clean air in the center is contaminated with room air entering from the borders of the ceiling. After passing the patient, the air flows toward the return air grilles near the floor (two-third of volume flow) and near the ceiling (one-third). The volume flow rate of supply air during measurements was 3301 m³/h, resulting in an air exchange rate of 13 changes per hour (according to the OR volume; Fig. 2A).

ORs at hospital II were provided with unidirectional LF supply air systems. This means that clean air flows straight down from the ceiling to the operating table, so that no mixing of existing air in the OR occurs on the way to the patient. The system was operated with 8000 m³/h of clean air, which results—according to the size of the air outlet of 3 m × 3 m—in an uniform flow velocity of 0.25 m/s (Fig. 2B).

Back to Top | Article Outline

Formula to Estimate the Mass of Inhaled Sevoflurane

The inhaled mass of sevoflurane was estimated according to the following equations:

  • 1. The inhaled mass (mVA) is the product of gaseous density (ρVA aer) and inhaled volume (VVA).

  • 2. The volume concentration of sevoflurane (cVA [ppm]) is defined as sevoflurane volume (VVA) divided by air volume (Vair).

  • 3. Density of gaseous sevoflurane.

  • 4. Temperature correction factor.

  • 5. Density of sevoflurane at 20°C.

Inhaled mass during time t: mVA = cVA × V × t × ρ VA aer.

The anesthesiologists’ respiratory minute volume (V = tidal volume × respiratory rate) was assumed to be 8 L/min.

Back to Top | Article Outline

Statistical Analysis

Acquisition of data started in March 2013. All patients who met the inclusion criteria (age 18–60 years, ASA physical status I or II, BMI 18–25 kg/m2, elective tonsillectomy, knee arthroscopy or neck surgery, and general balanced anesthesia with sevoflurane) were enrolled in this study. Exclusion criteria were ASA physical status ≥III, history of lung disease, increased risk of aspiration, known difficult airway, or the use of facemasks during surgery. The collection of data was stopped in September 2013, and analysis was performed. In 2014, additional 20 measurements were conducted.

Evaluations were performed using the programs Excel 2007 (Microsoft Corp., Redmond, WA) and IBM SPSS version 20 (IBM Corp., Armonk, NY) to calculate mean values, SD, as well as coefficients of variation. All data were tested for normal distribution by graphic models (histograms and Q-Q plots of residuals), as well as by using the Kolmogorov-Smirnov test and Lilliefors significance correction. For group differences, either the independent Student t test with unequal variances (residuals normally distributed) or the Wilcoxon-Mann-Whitney test (residuals not normally distributed) was used. The significance level was set as a 2-sided test, with an error probability <1% (P < 0.01).

Back to Top | Article Outline


Cross-Sensitivity Reactions

After 25 mL isopropyl alcohol, the measured sevoflurane concentration as reported by the gas monitor increased up to 25.51 ± 0.31 ppm. The maximal concentration was reached after (min:ss) 6:00 ± 1:19 minutes, whereas 50% of the maximal value could be detected after (min:ss) 32:10 ± 4:08 minutes and after (min:ss) 127:10 ± 16:05 minutes 10% remained (Fig. 3A). Larger quantities of isopropyl alcohol (30–50 mL) caused higher maximal sevoflurane concentrations with different time needed to reach 50% and 10% of the maximal concentration, although curve profiles looked similar (Fig. 3B).

Figure 3

Figure 3

N-propanol led to a maximal sevoflurane concentration of 4.15 ± 0.41 ppm, a reduction down to 50% was reached after (min:ss) 55:05 ± 13:07 minutes and 10% remained after (min:ss) 212:10 ± 46:14 minutes (Fig. 3C). For ethyl alcohol, the gas monitor detected a maximal sevoflurane concentration of 5.79 ± 0.38 ppm, which was reduced by 50% after (min:ss) 42:15 ± 9:05 minutes and reached 10% of the maximum after (min:ss) 172:18 ± 46:35 minutes (Fig. 3D).

The disinfectants Octeniderm and Cutasept, which contain isopropyl alcohol in different concentrations (Octeniderm 45 vol%, Cutasept 72 vol%), showed comparable curve profiles with maximal sevoflurane concentrations of approximately 32 ppm (Octeniderm 31.67 ± 6.39 ppm; Cutasept 32.24 ± 3.19 ppm). However, a decrease down to 50% or rather 10% of the maximal sevoflurane concentration lasted longer when using Cutasept than using Octeniderm ([min:ss] Cutasept: 82:06 ± 9:35 minutes, 363:07 ± 47:02 minutes; Octeniderm: 64:05 ± 24:23 minutes, 335:22 ± 66:19 minutes; Fig. 3, E and F).

The sevoflurane concentrations measured by the gas monitor neither corresponded to actual concentrations of alcohols or disinfectants (as the filter in the gas monitor was selectively calibrated for sevoflurane) nor to real sevoflurane concentrations (as sevoflurane was not used in this experimental setting). The absorption of light was measured at a single wavelength of 8.8 μm. Consequently, the signal resulting from the measurement cell is caused by the absorption of sevoflurane and by the absorption of any other gas in the measurement chamber at this single wavelength. Therefore, the resulting signal must be additive. Thus, the suspected cross-sensitivity reaction between the gas monitor (sevoflurane filter) and the alcohols/disinfectants can be confirmed.

All agents showed typical curve profiles, with regression lines equal to c = a × exp (−b × τ) {c: concentration [ppm], a: factor to calculate the regression, b: gradient, τ: delay time}. This means that cross-sensitivity reactions after skin disinfection in hospital can be calculated and excluded with reference to this equation because decay curves are continuous exponential functions without disturbances. Measured peak concentrations, which are higher than expected according to the described regression line, are induced by sevoflurane.

Back to Top | Article Outline

Sevoflurane Baseline

Sevoflurane concentration profiles in the anesthesiologists’ breathing zones always appeared similar with 6 characteristic peaks (n = 22; Fig. 4; Table 1; comp. Fig. 5).

Table 1

Table 1

Figure 4

Figure 4

Figure 5

Figure 5

Peak number 1 results from the first application of sevoflurane in the anesthesia induction room after inserting an airway device. The concentration decreases on the patient’s transfer to the OR, as there is a larger, sevoflurane-free air volume compared with the induction room. The second peak correlates with the time period when the patient gets positioned and connected to the respirator, and sevoflurane is used with high-flow oxygen to deepen anesthesia. The third peak arises once skin disinfection is conducted, representing cross-sensitivity reactions between disinfectants and sevoflurane. The following decline is mainly influenced by the decay rate of the disinfectant. After the disinfectant has subsided, only sevoflurane remains. The sevoflurane concentration typically stays constant and extends as a straight curve in parallel to the x-axis. The fourth peak results from coughing and extubation at the end of surgery. The sevoflurane concentration declines again on the patient’s transfer through the sevoflurane-free OR corridor and increases (fifth peak) when latencies occur. The sixth peak arises in the recovery room by the patient’s exhalation of remaining sevoflurane.

Back to Top | Article Outline

Induction of General Anesthesia

Forty patients underwent tonsillectomy. IV inductions were performed in 20 patients, whereas the other half obtained inhalational inductions. Three patients had to be excluded because accidental disconnection between the tracheal tube and the respirator occurred after sevoflurane was already used.

IV inductions lasted (min:ss) 8:04 ± 3:12 minutes on average. A first “sevoflurane peak” was caused by skin disinfection for placement of a venous catheter (cross-sensitivity reaction), before a second (real) peak appeared after tracheal intubation once sevoflurane was used to deepen anesthesia (fresh gas flow 2 L/min; Fig. 6A).

Figure 6

Figure 6

During IV inductions, anesthesiologists were exposed to

= 0.62 ± 0.33 ppm and

max = 1.23 ± 0.88 ppm, resulting in an estimated amount of 0.30 ± 0.22 mg inhaled sevoflurane per induction.

Inhalational inductions lasted (min:ss) 12:17 ± 5:20 minutes on average. First peaks were often caused by a leaking facemask, followed by a cross-sensitivity peak after skin disinfection to establish an IV catheter. The highest peak resulted from tracheal intubation (Fig. 6B).

During inhalational inductions, anesthesiologists were exposed to

= 2.43 ± 1.91 ppm (99% confidence level, 1.33–3.53 ppm), leading to 1.95 ± 1.54 mg inhaled sevoflurane per induction. In 5 patients, the anesthesiologist could smell sevoflurane during inhalational induction. Here, measured maximal sevoflurane concentrations were higher compared with inhalational inductions with subjectively airtight facemasks (smell of gas

= 43.08 ± 22.59 ppm versus no smell of gas

= 6.59 ± 4.25 ppm; P = 0.003).

Inhalational inductions caused higher average sevoflurane concentrations in the anesthesiologists’ breathing zones than IV inductions (

[ppm] INH 2.43 ±1.91 versus IV 0.62 ± 0.33; P < 0.001). Consequently, the inhaled amount of sevoflurane by anesthesiologists is also higher during inhalational compared with IV inductions (mVA [mg] INH 1.95 ± 1.54 versus IV 0.30 ± 0.22; P < 0.001).

Back to Top | Article Outline

Airway Management

Sevoflurane concentrations in the anesthesiologists’ breathing zones were analyzed in 45 patients undergoing knee arthroscopy (tube n = 22; LMA™ n = 23).

Two patients who obtained tracheal tubes were excluded because there were disconnections between the tube and the respirator. In 7 patients, the LMA™ was obviously not airtight (fresh gas loss of >150 mL/min detected by the respirator and smell of sevoflurane in the air noticed by the anesthesiologist). The anesthesiologists’ sevoflurane exposure in those patients was separately evaluated.

Mean sevoflurane concentrations were higher during surgery with LMATM (

[ppm] tube 0.37 ± 0.16 versus LMA™ 0.79 ± 0.53; P = 0.009). Maximal concentrations, as well as the estimated amount of inhaled sevoflurane, may also have been higher during surgery with LMA™ compared with tracheal tubes (

[ppm] tube 1.91 ± 0.91 versus LMATM 2.91 ± 1.81, P = 0.057; inhaled sevoflurane [mg/h] tube 1.47 ± 0.64 versus LMATM 2.73 ± 1.81, P = 0.019), although curve profiles looked similar in both groups (Fig. 7, A and B).

Figure 7

Figure 7

Patients with leaking LMATM (n = 7) may have had higher sevoflurane concentrations in the air compared with not leaking LMATM (

[ppm] LMATM leaking 2.39 ± 1.81 versus LMATM tight 0.79 ± 0.53, P = 0.013;

[ppm] LMA™ leaking 73.36 ± 69.12 versus LMATM tight 2.91 ± 1.81, P = 0.006; mVA [mg/h] LMATM leaking 9.58 ± 7.22 versus LMA™ tight 2.73 ± 1.81, P = 0.008). These generally higher sevoflurane concentrations can be measured during the entire operation time, although the gas leakage seems minimal and the patients’ventilation succeeds easily (Fig. 7C).

Furthermore, sevoflurane concentrations were measured in 10 patients with the gas sample line fixed at distances of 30 cm above the patients’ heads (tube n = 5; LMA™ n = 5).

Sevoflurane concentrations may have been higher in patients with LMATM compared with tracheal intubations (

[ppm] tube 0.38 ± 0.23 versus LMATM 0.87 ± 0.39, P = 0.042;

[ppm] tube 2.39 ± 1.21 versus LMA™ 4.80 ± 1.92, P = 0.045).

Back to Top | Article Outline

Airflow Technology

Narcoses of 24 adults, who underwent neck surgery either in an LF OR (n = 11) or in an OR with turbulent mixed flows (n = 13), were analyzed. The gas sample line was fixed within the individual breathing zone of the anesthesiologist (n = 14) or surgeon (n = 10).

Anesthesiologists’ sevoflurane exposure was low during all measurements, although trended higher in OR with turbulent mixed flows compared with supply air systems with LF (

[ppm] turbulent flow [TF] 0.29 ± 0.12 versus LF 0.13 ± 0.06, P = 0.012;

[ppm] TF 2.09 ± 1.24 versus LF 1.12 ± 0.72, P = 0.114; mVA [mg/h] TF 1.16 ± 0.50 versus LF 0.51 ± 0.25, P = 0.007).

The surgeons’ sevoflurane exposure may have been lower in OR with LF than in OR with turbulent mixed flows (

[ppm] TF 0.42 ± 0.19 versus LF 0.17 ± 0.04, P = 0.021;

[ppm] TF 1.35 ± 0.41 versus LF 0.97 ± 0.61, P = 0.281; mVA [mg/h] TF 1.69 ± 0.77 versus LF 0.69 ± 0.15, P = 0.021).

Back to Top | Article Outline


In this study, we demonstrated that each sevoflurane application—regardless of the type of induction, the chosen airway management, or the air flow technology used—caused measureable sevoflurane concentrations in the air. Consequently, anesthesia providers are exposed to sevoflurane during routine clinical work.

To date, a worldwide uniform time-weighted average exposure threshold limit concentration for sevoflurane has still not been defined. Some countries recommend exposure limits of 2.00 to 7.00 ppm for sevoflurane; however, most countries have not established any maximum admissible concentration for sevoflurane.6–8 Furthermore, we do not know whether chronic low-concentration sevoflurane exposure is toxic, and this debate remains controversial.1–5 Therefore, OR personnel should be aware of their individual occupational sevoflurane exposure.

First, we demonstrated the gasmonitor’s cross- sensitivity reaction between disinfectants and sevoflurane. This is of particular importance, as the usually high sevoflurane peak (>10.00 ppm) after skin disinfection is generated by the used disinfectant, as well as by sevoflurane, and can be detected for approximately 1 hour. This peak concentration as a function of time is the sum of emitted sevoflurane per volume unit of air and a time-dependent decay curve of the disinfectant, which is characterized by the maximal concentration and the decay factor in the exponent of the equation c = a × exp (−b × τ). We registered characteristic decay curves during our continuous measurements and were therefore able to calculate time-dependent sevoflurane concentrations only. As the gasmonitor has been used for years in anesthesia, given sevoflurane concentrations from previously published studies could be overestimated (10%–15%), depending on the kind and amount of used disinfectant.9–14

Second, we analyzed the anesthesiologists’ sevoflurane exposure during inhalational inductions and while LMATM were used. This is especially interesting because those situations are more often associated with higher waste anesthetic gas concentrations than IV inductions and tracheal intubations.

Sevoflurane exposure in the anesthesiologists’ breathing zones during inhalational induction was 3.53 ppm (99% upper confidence level;

= 2.43 ppm), which is comparable with findings from other studies (

= 15.91 ppm;

= 3.35 ppm).17,18 This is surprising because others used either a sevoflurane and nitrous oxide mixture or a sevoflurane in lower inspiratory concentrations (≤5% compared with 8% in our study), which both should lead to lower sevoflurane concentrations in the air.17,18 Reasons for this might be that we compensated the alcohols’ cross-sensitivity reaction and used an OR with LF air-conditioning system. In 5 patients, the anesthesiologist could smell sevoflurane during inhalational induction. This was a risk factor for high sevoflurane peak concentrations. It should be kept in mind that particularly leakages of the facemask result in higher maximal concentrations, as we used sevoflurane with 8% inspiratory concentration and with a fresh gas flow of 10 L/min. However, although high peak concentrations might occur during inhalational induction, the mean sevoflurane concentration is consistently low. We conclude that inhalational inductions are accompanied by higher sevoflurane exposure compared with IV inductions, although differences were lower than expected.

We further demonstrated that sevoflurane concentrations in the anesthesiologists’ breathing zones, as well as in the air, were higher while LMA™ were used compared with tracheal intubation. Basically, these results are consistent with figures from other studies (LMA™ 1.00 ppm), although a sevoflurane-nitrous oxide mixture was used by others (lower concentration expected).19 Again, the cross-sensitivity reaction might have led to these higher sevoflurane concentrations. Moreover, it should be noted that approximately 30% of our LMA™ were leaking, accompanied by a trend for a higher sevoflurane exposure compared with obviously “tight” LMA™ (fresh gas loss <150 mL/min and no smell of sevoflurane). Those figures are possibly too high, as only 15% of LMA™ were leaking in other studies.20,21 However, we observed that a leaking LMA™ can be replaced by another kind of LMA™ (e.g., supraglottic airway I-gel® with noninflatable cuff [Intersurgical Ltd., Berkshire, UK]), and a tube must not necessarily be used.

In addition, we tested the protective effects of OR airflow systems during surgery. As expected, average sevoflurane concentrations were lower in OR with LF compared with turbulent mixed flow. Interestingly, overall mean sevoflurane concentrations were lower than described by other investigators, although they used either a sevoflurane-nitrous oxide mixture (sevoflurane 0.60 ppm) or a different gas detection method without cross-sensitivity reaction (sevoflurane 1.14 ppm).22,23 This could partly be explained by the fact that most of the patient’s face and the tube are covered with sterile drapes during neck surgery. In addition, the anesthesiologist is placed at the patient’s side and not close to the head, where sevoflurane concentrations might be higher. Therefore, sevoflurane concentrations were measured in the surgeons’ breathing zones during surgery. Concentrations were marginally higher than in the anesthesiologists’ breathing zones, however, still lower as previously described. We conclude that LF air-conditioning systems reduce sevoflurane concentrations more effectively than TF air-conditioning systems because differences reach up to 55% (anesthesiologists’ breathing zones) and 59% (surgeons’ breathing zones). This is impressive, because situations with usually higher sevoflurane exposure (patient’s positioning, skin disinfection, extubation) were already excluded. Furthermore, other beneficial effects of LF supply air systems, such as a lower germ number in the operating area and lower germ sedimentation on the instrument table, which reduces the risk of postoperative wound infections, should be kept in mind.24

As clinical work is heterogeneous, depending on the patient, the kind of surgery, and clinical standards, it is not possible to give detailed information about the amount of sevoflurane anesthesiologists inhale during a workday. In this study, the mean sevoflurane concentration during anesthesia was <2.00 ppm. Exceptions were only inhalational inductions and leaking LMA™, although even during these particular circumstances mean sevoflurane concentrations were just 3.53 and 4.21 ppm (99% upper confidence levels). Furthermore, the amount of inhaled sevoflurane per day was always <10.00 mg. Indeed, it should be mentioned that the amount of inhaled sevoflurane varies with the anesthesiologists’ respiratory minute volume, which depends on sex, body size, metabolic rate, pulmonary function, and grade of activity. Here, estimates were performed with a respiratory minute volume of 8 L/min. However, we assume that the values given may vary up to 30%.

In our hospital, inhalational inductions are always performed in OR with LF. We do not use obviously leaking LMA™, although the patient’s ventilation succeeds easily and we are aware of our daily occupational chronic sevoflurane exposure. And although measured sevoflurane concentrations were lower than expected, further efforts must be taken to minimize occupational sevoflurane exposure.

Back to Top | Article Outline


Name: Jennifer Herzog-Niescery, MD.

Contribution: This author created the study design, participated in the clinical part of this study, collected and analyzed data, and wrote the manuscript.

Attestation: Jennifer Herzog-Niescery approves the final manuscript and attests to the integrity of the original data and the analysis reported in this manuscript and is also the archival author.

Name: Nikolaj Matthias Botteck, MD.

Contribution: This author helped with the study design, collected clinical data, and prepared the manuscript.

Attestation: Nikolaj Matthias Botteck approved the final manuscript and attests to the integrity of the original data and the analysis reported in this manuscript.

Name: Heike Vogelsang, MD.

Contribution: This author helped conduct the study in St. Josef Hospital, collect data, and prepare the manuscript.

Attestation: Heike Vogelsang approved the final manuscript.

Name: Philipp Gude, MD.

Contribution: This author helped conduct the study in St. Elisabeth Hospital, collect data, and prepare the manuscript.

Attestation: Philipp Gude approved the final manuscript.

Name: Horst Bartz, Dr-Ing.

Contribution: This author helped design the study, collect and analyze data from experimental tests, and prepare the manuscript.

Attestation: Horst Bartz approved the final manuscript and attests to the integrity of the original data and the analysis reported in this manuscript.

Name: Thomas Peter Weber, MD.

Contribution: This author helped design the study, analyze data, and prepare the manuscript.

Attestation: Thomas Peter Weber approved the final manuscript and attests to the integrity of the original data and the analysis reported in this manuscript.

Name: Hans-Martin Seipp, MD, Dipl-Ing.

Contribution: This author helped design the study, collect and analyze data from experimental tests, and prepare the manuscript.

Attestation: Hans-Martin Seipp approved the final manuscript and attests to the integrity of the original data and the analysis reported in this manuscript.

This manuscript was handled by: Sorin J. Brull, MD, FCARCSI (Hon).

Back to Top | Article Outline


1. Baden JM, Kelley M, Mazze RI. Mutagenicity of experimental inhalational anesthetic agents: sevoflurane, synthane, dioxychlorane, and dioxyflurane. Anesthesiology. 1982;56:462–3
2. Eger EI II, Laster MJ, Winegar R, Han C, Gong D. Compound A induces sister chromatid exchanges in Chinese hamster ovary cells. Anesthesiology. 1997;86:918–22
3. Morio M, Fujii K, Satoh N, Imai M, Kawakami U, Mizuno T, Kawai Y, Ogasawara Y, Tamura T, Negishi A. Reaction of sevoflurane and its degradation products with soda lime. Toxicity of the byproducts. Anesthesiology. 1992;77:1155–64
4. Wiesner G, Schiewe-Langgartner F, Lindner R, Gruber M. Increased formation of sister chromatid exchanges, but not of micronuclei, in anaesthetists exposed to low levels of sevoflurane. Anaesthesia. 2008;63:861–4
5. Teschke K, Abanto Z, Arbour L, Beking K, Chow Y, Gallagher RP, Jong B, Le ND, Ratner PA, Spinelli JJ, Dimich-Ward H. Exposure to anesthetic gases and congenital anomalies in offspring of female registered nurses. Am J Ind Med. 2011;54:118–27
6. National Institute of Occupational Safety and Health (NIOSH). NIOSH Pocket Guide to Chemical Hazards. 1994 Washington, DC Government Printing Office
7. National Institute of Health Office of Research Services. Division of Occupational Health and Safety: Waste Anesthetic Gas (WAG) Surveillance Program. 2014 Technical Assistence Branch
8. Central Institute for Labour Protection – National Research Institute (CIOP-PIB). Principles and Methods of Assessing the Working Environment. 2007(No 3) Warszawa, Poland Sevoflurane
9. Hoerauf K, Funk W, Harth M, Hobbhahn J. Occupational exposure to sevoflurane, halothane and nitrous oxide during paediatric anaesthesia. Waste gas exposure during paediatric anaesthesia. Anaesthesia. 1997;52:215–9
10. Hoerauf KH, Hartmann T, Zavrski A, Adel S, Burger HJ, Koinig H, Zimpfer M. Occupational exposure to sevoflurane during sedation of adult patients. Int Arch Occup Environ Health. 1999;72:174–7
11. Byhahn C, Heller K, Lischke V, Westphal K. Surgeon’s occupational exposure to nitrous oxide and sevoflurane during pediatric surgery. World J Surg. 2001;25:1109–12
12. Schebesta K, Lorenz V, Schebesta EM, Hörauf K, Gruber M, Kimberger O, Chiari A, Frass M, Krafft P. Exposure to anaesthetic trace gases during general anaesthesia: CobraPLA vs. LMA classic. Acta Anaesthesiol Scand. 2010;54:848–54
13. Blokker-Veldhuis MJ, Rutten PM, De Hert SG. Occupational exposure to sevoflurane during cardiopulmonary bypass. Perfusion. 2011;26:383–9
14. Versichelen L, Rolly G, Vermeulen H. Accumulation of foreign gases during closed-system anaesthesia. Br J Anaesth. 1996;76:668–72
15. Walder B, Lauber R, Zbinden AM. Accuracy and cross-sensitivity of 10 different anesthetic gas monitors. J Clin Monit. 1993;9:364–73
16. Scott CC. Laminar-linear flow system of ventilation. Its application to medicine and surgery. Lancet. 1970;1:989–93
17. Byhahn C, Strouhal U, Westphal K. [Exposure of anesthetists to sevoflurane and nitrous oxide during inhalation anesthesia induction in pediatric anesthesia]. Anaesthesiol Reanim. 2000;25:12–6
18. Hasei M, Hirata T, Nishihara H, Tanigami H, Takashina M, Mori T. Occupational exposure of operating room staff to anesthetic gases during inhaled induction—a comparison with intravenous anesthesia induction. Masui. 2003;52:394–8
19. Gustorff B, Lorenzl N, Aram L, Krenn CG, Jobst BP, Hoerauf KH. Environmental monitoring of sevoflurane and nitrous oxide using the cuffed oropharyngeal airway. Anesth Analg. 2002;94:1244–8
20. Hönemann CW, Hahnenkamp K, Möllhoff T, Baum JA. Minimal-flow anaesthesia with controlled ventilation: comparison between laryngeal mask airway and endotracheal tube. Eur J Anaesthesiol. 2001;18:458–66
21. Fröhlich D, Schwall B, Funk W, Hobbhahn J. Laryngeal mask airway and uncuffed tracheal tubes are equally effective for low flow or closed system anaesthesia in children. Br J Anaesth. 1997;79:289–92
22. Hoerauf KH, Hartmann T, Acimovic S, Kopp A, Wiesner G, Gustorff B, Jellinek H, Krafft P. Waste gas exposure to sevoflurane and nitrous oxide during anaesthesia using the oesophageal-tracheal Combitube small adult. Br J Anaesth. 2001;86:124–6
23. Tankó B, Molnár C, Budi T, Peto C, Novák L, Fülesdi B. The relative exposure of the operating room staff to sevoflurane during intracerebral surgery. Anesth Analg. 2009;109:1187–92
24. Hirsch T, Hubert H, Fischer S, Lahmer A, Lehnhardt M, Steinau HU, Steinstraesser L, Seipp HM. Bacterial burden in the operating room: impact of airflow systems. Am J Infect Control. 2012;40:e228–32
© 2015 International Anesthesia Research Society