AYOUB ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 85, NO. 6,

Transcription

1 AYOUB ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 85, NO. 6, SPECIAL REPORT Determination of Low-Level Residual Ethylene Oxide by Using Solid-Phase Microextraction and Gas Chromatography KAMAL AYOUB, LEONARD HARRIS, and BILL THOMPSON Tyco/Healthcare-Kendall, 2010 E. International Speedway Blvd, DeLand, FL Current methods of analysis for ethylene oxide (EO) in medical devices include headspace and simulated-use extractions followed by gas chromatography with either a packed or a capillary column. The quantitation limits are about g/g for a packed column and about g/g for a capillary column. The current allowable levels of EO on medical devices sterilized with EO gas as outlined in International Organization for Standardization (ISO) may be significantly reduced from current levels by applying the ISO Draft International Standard method for establishing allowable limits. This may require EO test methods with detection and quantitation limits that are much lower than those of the currently available methods. This paper describes a new method that was developed for the determination of low-level EO by solid-phase microextraction using the direct-immersion method. Factors such as temperature and stirring were found to affect absorption efficiency and absorption time. A low extraction temperature (about 6 C) was found to be more efficient than room-temperature extraction. Stirring was found to reduce absorption time by about 50%. Under these conditions, detection and quantitation limits of and g/g, respectively, were obtained by using a capillary column. As a result, this method makes compliance with lower EO limits feasible. Solid-phase microextraction (SPME) is a simple and inexpensive sample extraction technique that allows lower detection limits for trace organic compounds. It was developed in the early 1990s at the University of Waterloo in Ontario, Canada (1 3). The technique became commercially available in 1993 with the introduction of the 100 µm polydimethylsiloxane (PDMS)-coated fiber (4). Currently, many different fiber coatings are available for a wide range of applications. All manufactured SPME fibers contain a fused-silica core coated with the desired phase. The technique involves the exposure of a chemically coated microfiber Received April 12, Accepted by JS July 12, Corresponding author s kamal.ayoub@tycohealthcare.com. to a gaseous or liquid sample or the headspace above the sample. Analytes come to equilibrium with the fiber according to their affinity for the solid phase. The fiber is then directly injected into the gas chromatograph without solvent. The published literature contains several hundred papers on the application of SPME to the determination of organic analytes. The headspace technique is usually used for volatile analytes (5), whereas the direct-immersion technique (directly dipping the fiber into the liquid sample) is used for less volatile analytes (6). In this paper, we report a method for the use of SPME in the determination of ethylene oxide (EO). EO is a colorless, flammable gas at ordinary room temperature and pressure. It has many uses, such as fumigation of foods and textiles and sterilization of medical devices (7). EO is commonly used to sterilize medical devices that are not compatible with sterilization by steam, dry heat, or irradiation. EO is toxic not only to microorganisms but also to patients and healthcare workers exposed to EO as residues in medical devices, and the toxicity of EO includes carcinogenicity and mutagenicity. As a result, the U.S. Food and Drug Administration (FDA) had proposed limits for EO residues on medical devices in a proposed rule published in 1978 (8). The International Organization for Standardization (ISO) developed ISO , Biological Evaluation of Medical Devices Part 7: Ethylene Oxide Sterilization Residuals, to specify the requirements for establishing allowable limits for EO residues on medical devices. In 1998, a technical information report (TIR-19) was published by the Association for the Advancement of Medical Instrumentation (AAMI) to provide guidance for the application of ISO (9, 10). The FDA recognized ISO and AAMI TIR-19 (11) as an acceptable means of meeting the agency s requirements concerning EO residuals. Current methods of analysis for EO residuals from medical devices include headspace and simulated-use extractions. In the headspace technique, a sample is placed in an extraction vial, and the vial is heated. The headspace above the sample is sampled and analyzed by gas chromatography (GC). In the simulated-use extraction technique, a representative sample is extracted with a solvent such as water, and the extract is analyzed by GC. Both techniques have quantitation limits of about 0.1 µg/g for a capillary column and 0.5 µg/g for a packed column. The current allowable levels of EO as outlined in ISO (Table 1) may be significantly reduced as a result of

2 1206 AYOUB ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 85, NO. 6, 2002 Table 1. EO limits outlined in ISO a Table 2. GC method parameters Exposure category EO limit Parameter Packed column Capillary column Limited Prolonged Permanent Intraocular lenses Apheresis kits Blood oxygenator 20 mg per device 60 mg maximum (20 mg in first day, 60 mg in first month) 2.5 g maximum (20 mg in first day, 60 mg in first month, and 0.1 mg per day average) 0.5 µg per lens per day, 1.25 µg total 15 mg maximum 60 mg maximum (20 mg in first day) Oven temperature 85 C 30 C (initial) for 2 min, 30 C/min to 100 C, hold for 1.5 min Injector temperature 200 C 225 C Detector temperature 250 C 225 C Helium flow 35 ml/min At 10 psi Injection volume N/A N/A a AAMI TIR-19 contains an additional requirement of 250 µg/g for small devices, provided that the total mg per device is not exceeded in a given category. applying the ISO Draft International Standard (DIS) method for establishing allowable limits. This will require EO test methods with detection and quantitation limits that are much lower than those of the currently available methods. In this paper, we discuss the application of SPME as a new technique to obtain detection limits that are much lower than those of current methods, thus making compliance with lower EO limits feasible. Experimental Reagents EO standard of 99.9% purity was purchased from Supelco (Bellefonte, PA) and used without further purification. Preparation of Standard and Test Solutions A 500 µg/g stock standard was prepared by dilution with deionized water. A series of working standard solutions (0.01, , 0.5, 1, and 2 µg/g) was prepared by dilution with deionized water. The working standards were stored in a refrigerator at 2 8 C. Additional standards of known concentrations were also prepared and treated as samples. Instrumentation Four different SPME fibers were obtained from Supelco and evaluated to determine the most suitable for the EO extraction. The fibers evaluated were PDMS/divinylbenzene (PDMS/DVB), Carbowax/DVB (CW/DVB), Carboxen /PDMS (CAR/PDMS), and DVB/Carboxen TM /PDMS (CAR/PDMS/DVB). Before the first use, fibers were conditioned as recommended by the manufacturer by heating them in the injection port of the chromatographic system for h at C, depending on the fiber coating. Analysis was performed with a Perkin-Elmer Autosystem (Norwalk, CT) gas chromatograph equipped with a flame ionization detector. A packed column (3% Carbowax 20M on Chromosorb 101, 80/100 mesh; Supelco) and a fused-silica capillary column (Crossbond trifluoropropylmethyl polysiloxane, Restek RTX -200) were used as separate columns. The GC method parameters for both columns are given in Table 2. Data were gathered with Turbochrom software, which was provided with the equipment. Figure 1. Effects of SPME time on the detector response of the analyte, obtained by using a packed column. The analysis was performed in the liquid phase at room temperature without agitation, using the 1 g/g EO standard and the CAR/PDMS fiber.

3 AYOUB ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 85, NO. 6, Figure 2. Effects of extraction temperature on absorption efficiency, obtained by using the CAR/PDMS fiber and a packed column. SPME Procedure The affinity of each fiber for EO was initially evaluated by direct sampling of the aqueous phase. This was achieved by immersion of the fiber half way in a 10 ml vial containing a known EO standard. The SPME device was clamped in place for 30 min at room temperature without agitation. It was found that the highest affinity was with the CAR/PDMS fiber. Once the most efficient and adsorbent fiber was identified, procedures were performed to optimize extraction time, extraction temperature, and agitation. Low-temperature extractions were performed by placing the extraction samples in a temperature-monitored ice bath at 6 C. Agitation was performed by using a magnetic stir bar. Results and Discussion Effects of Extraction Time on the Analytical Signal A series of experiments was conducted to investigate the effects of extraction time on the adsorption of EO onto the optimal fiber. The resultant absorption extraction time plot, obtained by introducing the fiber into the liquid phase of the 1 µg/g EO standard in a 10 ml vial without agitation, for durations of 2 90 min, is shown in Figure 1. The results show that the system reached equilibrium after the fiber was exposed to the liquid phase for an extraction time of approximately min without agitation. Effects of Extraction Temperature on the Analytical Signal Extractions were performed at room temperature (20 C) and at refrigerator temperature (6 C) by using the 1 µg/g standard in a 10 ml vial without agitation. The resulting plot (Figure 2) shows that the lower extraction temperature resulted in increased efficiency. This is most likely attributed to the reduced kinetic energy of the EO molecules and thus fewer transitions between the active sites of the fiber and the liquid phase. Figure 2 also shows that the equilibrium times were similar for extraction at the lower temperature and at room temperature. The use of headspace extraction at 37 and 50 C was also evaluated. The samples were placed in an oven at the desired temperature with the SPME fiber for 1 h. After 1 h, the samples were analyzed for EO as in previous testing. The area Figure 3. column. Effects of agitation on EO extraction time at 6 C, obtained by using the CAR/PDMS fiber and a capillary

4 1208 AYOUB ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 85, NO. 6, 2002 Table 3. MDL and MQL values obtained for the capillary column by extraction at 6 C with stirring MDL and MQL Std value, µg/g a Calc. std value, µg/g Sample value, µg/g Calc. sample value, µg/g Parameter Value Average, µg/g Accuracy, % SD b RSD, % c No. of runs t d MDL (SD t), µg/g MQL (MDL 5), µg/g a Std = standard. b SD = standard deviation. c RSD = relative standard deviation. d At n 1 and α = count for EO was then compared with the values previously obtained by using the direct-injection method. It was found that the detector response for the headspace extraction at both elevated temperatures (37 and 50 C) was about half of that observed by using the direct-immersion method for extraction. Effects of Agitation on the Analytical Signal A magnetic stirrer was used to gently agitate the solution. Agitation of the sample was found to reduce the time needed to reach equilibrium by 50% (reduction from 20 to 10 min) as shown in Figure 3. Effects of Extraction Vial Size on the Analytical Signal Two vial sizes were examined to evaluate the effects of sample size on the amount of EO extracted. Filled vial sizes of 10 and 2 ml containing 1 µg/g EO were evaluated. The results showed no difference in the GC signals for the 2 sample sizes. As a result, it was determined that a 2 ml sample size is sufficient. Method Detection Limit Studies Calibration curves were obtained for both packed and capillary columns by using single injections of prepared standards at various levels. The optimum conditions were used as outlined above for each column. For both packed and capillary columns, the correlation coefficient was >99.9%. The calibration curves were then used to determine the method detection limit (MDL) and the method quantitation limit (MQL) for both columns. The MDL and MQL were calculated by using the methodology given in standard methods (12). Tables 3 and 4 show the MDL and MQL for the cap- Table 4. MDL and MQL values obtained for the packed column by extraction at 6 C with stirring MDL and MQL Std value, µg/g a Calc. std value, µg/g Sample value, µg/g Calc. sample value, µg/g Parameter Value Average, µg/g Accuracy, % SD b RSD, % c No. of runs t d MDL (SD t), µg/g 0.01 MQL (MDL 5), µg/g 0.05 a Std = standard. b SD = standard deviation. c RSD = relative standard deviation. d At n 1 and α = 0.01.

5 AYOUB ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 85, NO. 6, illary column and the packed column, respectively, for direct immersion in solvent (water) at 6 C, with stirring. The packed column was found to have values of and µg/g for the MDL and the MQL, respectively, whereas the values for the capillary column were and µg/g, respectively. Both methods have similar accuracy and repeatability, as shown in Tables 3 and 4. Conclusions This paper demonstrates that SPME can be used to determine low-level EO by the direct-immersion method. Factors such as temperature and stirring were found to affect absorption efficiency and absorption time. Low-temperature extraction (approximately 6 C) was found to be more efficient than room-temperature extraction. Stirring was found to reduce absorption time by about 50%. The size of the extraction vial (2 or 10 ml) had no impact on the amount of EO extracted. By using a capillary column, detection and quantitation limits of and 0.01 µg/g, respectively, were obtained. As a result, this method makes compliance with lower EO limits feasible. Acknowledgments We express our sincerest thanks to Bob Shirey (Supelco) for his support and for providing the SPME fibers and equipment needed to perform this research. References (1) Belardi, R.P., & Pawliszyn, J. (1989) J. Water Pollut. Res. Can. 24, 179 (2) Arthur, C.L., & Pawliszyn, J. (1990) Anal. Chem. 62, 2145 (3) Arthur, C.L., Pratt, K., Motlagh, S., Pawliszyn, J., & Belardi, R.P. (1992) J. High Resolut. Chromatogr. 15, 741 (4) Shirey, R.E. (2000) J. Chromatogr. Sci. 38, 109 (5) Bartelt, R.J. (1997) Anal. Chem. 69, 364 (6) Hernandez, F., Beltran, J., Lopez, F.J., & Gasper, J.V. (2000) Anal. Chem. 72, 2313 (7) The Merck Index (1989) 11th Ed., Merck & Co., Inc., Rahway, NJ (8) U.S. Food and Drug Administration (1978) Fed. Regist. 43, 122 (9) Cyre, W.H., & Page, B.F.J. (2000) Med. Device Diagn. Ind. February, (10) Centola, D.T., Ayoub, K.I., Lao, N.T., Lu, H.T.C., & Page, B.F.J. (2001) J. AOAC Int. 84, (11) Guidance for ANSI/AAMI/ISO :1995, Biological Evaluation of Medical Devices Part 7:1995 (1999) AAMI TIR-19 (Amendment 1), Association for the Advancement of Medical Instrumentation, Arlington, VA (12) Standard Methods for the Examination of Water and Waste Water (1992) 18th Ed., American Public Health Association, American Waterworks Association, and Water Environment Federation, Washington, DC