Ethylene Oxide Gaseous Sterilization I. Concentration and Temperature Effects ROBERT R. ERNST

AND

JAMES J. SHULL

Wilniot Castle Company, Rochester, New York Iteceived for publication February 7, 1962

cal sterilization procedures has involved extension of these conclusions of Phillips into temperature and concentration ranges for which the basic time factors had not been established. As a result, some of the ethylene oxide use conditions have been arbitrarily chosen without accurate experimental support (Znamirowski, McDonald, and Itoy, 1960). During the course of our investigations of ethylene oxide sterilization at temperatures above 37 C, it became evident that the relationships of concentration, temperature, and time were not as simple as reported. It was the purpose of the present study to review the work of Phillips (1949), and extend the temperatures investigated to include those commonly involved in use of commercially available ethylene oxide sterilizers. For convenience and consistency of nomenclature, the authors use the term "thermochemical death time" (TCDT) to describe the minimal time for complete loss of viability of a microbial population upon exposure to a chemical agent or agents under specified conditions of concentration and temperature. In the system under study, ethylene oxide is considered the primary chemical agent, with water vapor contributing to the reaction.

ABSTRACT

ERNST, ROBERT R. (Wilmot Castle Co., Rochester, N.Y.) AND JAMES J. SHULL. Ethylene oxide gaseous sterilization. I. Concentration and temperature effects. Appl. MIicrobiol. 10:337-341. 1962.-The relationships of reaction temperature and concentration of gaseous ethylene oxide to the time required for inactivation of air-dried Bacillus subtilis var. niger spores are more complex than previously reported. A plot of temperature vs. the logarithm of "thermochemical death time" (TCDT) resulted in a straight line between 18 and 57 C for systems of "high" ethylene oxide concentration. The TCDT values were independent of ethylene oxide concentrations above certain temperature-dependent limits. A given ethylene oxide concentration produced a TCDT curve identical in the upper temperature regions with that for higher concentrations. As the temperature was lowered beyond a critical point, this curve diverged from that for higher concentrations, as a straight line of lesser slope. Thus, a series of curves exists for a range of ethylene oxide concentrations. They are characterized by two segments, both logarithmic, intersecting at a critical temperature for each concentration. The intersecting point is at a temperature inversely related to the ethylene oxide gas MATERIALS AND METHODS concentration. The temperature quotient for the high Spore preparations. Bacillus subtilis var. niger was chosen temperature segments of all systems was 1.8. This value as the test organism for ease of handling, stability of its was characteristic for ethylene oxide concentrations of in dry preparations, and because of the relatively spores 440 and 880 mg/liter at temperatures above 40.6 and of its spores to ethylene oxide (Phillips, resistance high 33.4 C, respectively. Below these critical temperatures, Phillips, 1949; Friedl, Ortenzio, and and Kaye 1949; the Qlo values for the respective systems were 3.2 and 2.3. Stuart, 1956). The strain used in this investigation is a descendant of B. globigii, 356-S.C. no. 4, N. R. Smith Gaseous ethylene oxide has been gaining increased use strain (Smith, Gordon, and Clark, 1952). Heat-shocked spores of the test organism were used as as a sterilizing agent in the medical and pharmaceutical fields since the appearance of a review and series of papers inocula for starter cultures in a casein acid digest medium on the subject by Phillips and Kaye (1949), Phillips of the following composition: casein acid hydrolyzate (1949), Kaye (1949), and Kaye and Phillips (1949). By powder (General Biochemicals, Inc., Chagrin Falls, Ohio), far the greatest stimulus to its use, and the primary source 10.0 g; Amber 300 yeast extract (Amber Laboratories, of reference regarding the process, has been this set of Inc., Milwaukee, Wis.), 5.0 g; glucose, 5.0 g; K2HPO4, papers. Phillips (1949) concluded that the Qlo of the re- 5.0 g; CaCl2 2H20, 0.066 g; MnSO4.H20, 0.03 g; mineral action system was 2.74 within the range from 5 to 37 C, solution, 1 ml; and tap water, 1,000 ml. The mineral soand that each doubling of the ethylene oxide concentration lution consisted of thiamine HCl, 0.25 g; MgSO4 .7H20, approximately halved the time required for sterilization, 0.5 g; FeSO4 7H20, 0.1 g; and water, 10 ml. The medium in the range from 22.1 to 884 mg/liter. It is somewhat was filtered, adjusted to pH 7.0, and sterilized in steain unfortunate that application of ethylene oxide to practi- at 121 C for 20 min. 3,37

338

ERNST AND SHULL

After shaking for 24 hr at 32 C, the starter culture was used to inoculate large flasks of the same medium. These were incubated with shaking for 4 days at 32 C, and the sporulated culture was placed at 45 C overnight to allow autolysis of vegetative cells. The spores were then harvested by centrifugation, washed three times in distilled water, and heat-shocked at 80 C for 10 min. Reddish (1957) recommended use of unheated spores for testing chemical sterilants, since heating often increases susceptibility of the spores to the chemical under test. Heat-shocking was employed in the present study to insure a high percentage of germination and outgrowth of viable spores in the recovery medium. The validity of this procedure was established by unpublished work in this laboratory; no significant differences in the rate of destruction of heat-shocked and unheated spores of B. subtilis var. niger by gaseous ethylene oxide were found. Carrier and distribution. Spores on glass or steel appear to be less susceptible to the action of ethylene oxide than are spores placed on paper or cotton fiber (Phillips, 1949). This fact was re-established in our laboratory. The observed differences in susceptibility might be related to a combination of three factors: adsorption of water vapor, adsorption of ethylene oxide, and exposed surface area. We deemed it desirable to avoid materials which might significantly influence the concentration of ethylene oxide or water in the immediate vicinity of the spore, and therefore chose glass in preference to fibrous materials as the carrier for spore preparations. Glass also more closely approximates materials processed in ethylene oxide sterilizers, and subjects the system to more rigorous conditions. Heat-shocked spores in a 10-ml distilled-water suspension were placed in a 300-ml polyethylene bottle with approximately 4,000 glass beads (4 mm diam). The bottle was rotated on its side at 30 rev/min while dry air at 45 C was forced into the bottle through a hole in the bottom. This permitted uniform deposition and drying of spores on the beads. The production of larger batches resulted in gross population irregularities from bead to bead and results obtained with such preparations showed serious anomalies. Since population numbers and distribution varied from batch to batch, direct comparisons were not made among restults obtained with different preparations. Plate counts of the bead preparations were performed in accordance with Standard Mlethods for the Examination of Dairy 1'roducts (11th ed.) except that distilled water was used as the diluting fluid. Satisfactory dispersion of the spores was accomplished by shaking the beads in the first dilution blank. Experimental procedure. For determination of the TCDT at various ethylene oxide concentrations and temperatures, 30 spore-coated glass beads were used per test. Three beads were placed in each tube to minimize the effect of variations in count, and a set of ten tubes was

[VOL. 10

placed in the chamber shown in schematic form in Fig. 1. With metal drop-caps in place the tubes were allowed to equilibrate to the desired temperature. When the temperature remained constant for 10 min, the cycle was started. The chamber was first evacuated to 40 mm Hg absolute pressure through valve (A). Valve (A) was closed, and with the heating tape turned on, a measured volume of water was injected through septum (B) to produce an estimated 40 % relative humidity in the chamber at the chosen temperature (Kaye, 1949; Phillips, 1949, 1961). The system was allowed to equilibrate for 5 min; a measured volume of liquid ethylene oxide was admitted through valve (C) and the hot piping to the chamber; and timing was begun. To stop the reaction, ethylene oxide was removed by evacuating the chamber through valve (A). Valve (A) was then closed, and filtered air was admitted through valve (D). The samples were immediately transferred to Trypticase soy broth (Baltimore Biological Laboratory, Baltimore, Md.) and incubated for 7 days at 32 C. Lack of visual evidence of growth in the tubes was interpreted as indication of sterility of the beads. Gas concentrations were established by volumetric delivery of liquid ethylene oxide to the reaction system, the volume of the chamber being known. This procedure was found to be accurate within 1 % when compared to partialpressure measurements and gas-chromatographic analysis. Chromatographic determinations were made with a Vapor Fractometer (Perkin-Elmer Corp., Norwalk, Conn.), model 154, using Perkin-Elmer Column W, a two-meter column of polyethylene glycol 1500 on Teflon (E. I. duPont de Nemours Inc., Wilmington, Del.). Fractionations were made at 100 C with helium as the carrier gas. Ethylene oxide, water vapor, and air (including C02) were determined quantitatively by peak-height analysis.

-THERMOCOUPLE

h \

JACKET WATER

CONTAMINATED BEADS T __THERMISTOR CONTROLLER

FIG. 1. Diagrammatic representation of the cylindrical pressure vessel and its accessory fittings, showing position of beads, tubes, and

thermocoutple.

ETHYLENE OXIDE CONCENTRATION AN\-D TEMWPERATURE RESULTS

The results reported were obtained with a single batch of spore-coated beads. Similar results, differing only in absolute time values, were obtained with two previous and one subsequent set of beads. Figure 2 is a histogram showing the distribution of count of individual beads, as determined by standard plate count. The observed overall range of counts covered 1.7 logs with a mean of 2.5 X 105 spores per bead. Table 1 lists the results from the total of 160 individual experiments. Each experiment involved exposure of a set of ten tubes containing three beads each to a specific ethylene oxide concentration, temperature, and time. WVhen the exposed beads were tested for sterility, the restults were placed in one of three categories: (i) beads from all ttubes were sterile; (ii) viable spores were recovered from each tube; or (iii) viable spores were recovered from a fraction of the ten tubes, hereafter referred to as a "mixed set." Those results falling within the first two categories, MEAN=2.5+0.12X 1O5 SPORES PER BEAD

121 10 8

z

2

5.0

4.6

5.4 5.8 LOG NUMBER OF SPORES PER BEAD

6.2

FIG. 2. Distributtion of spores of Bacilluis subtilis var. niger dried on glass beads. TABLE 1. Sterility test results from spore preparations* exposed to ethylene oxide gas Combinationst

No. borderline

Unmixed sets 0/10 or 10/10

Mixed sets 1/10 2/10 3/10 4/10 5/10

6/10, 7/10, 8/10, 9/10 Subtotal Total

Sets observed

125 2

5 1 2 6 2

7 6 1 3 8 10 35 160

* Bacillus subtilis var. niger on glass beads. t Number of tubes within a set of ten showing growth in Trypticase soy broth at 32 C in 7 days.

339

representing 78 %/ of the total, were used to establish the curves in Fig. 3. Separate curves between 18 and 57 C were established for each concentration used. With time of exposure and temperature of each experiment as coordinates, points were established on a graph and marked to indicate the results obtained from the sterility tests. When a sufficient number of points had been established to permit reasonable accuracy, a curve was drawn to indicate the locus of demarcation between sterile and nonsterile sets after exposure. The curves so derived are considered the TCDT curves for the various systems. The TCDT curves for systems with ethylene oxide concentrations of 440 and 880 mg/liter exhibited two logarithmic segments. In the higher temperature range, the curves for the two concentrations were essentially identical. As the temperature decreased, a break was observed first in the curve for 440 mg/liter at 40.6 C and then in the curve for 880 mg/liter at 33.4 C, resulting in two nonparallel curves of lesser slopes. Extension of the high-temperature segments of the curves for 440 and 880 mg/liter in the direction of lower temperature then provided a hypothetical "minimal TCDT" curve for the temperature range from 18 to 57 C. The validity of the extrapolated curve was tested using an ethylene oxide concentration of 1,500 mg/liter. The results confirmed the extrapolation (Fig. 3C). Thus, the system exhibited a maximal reaction rate for each temperature. When the concentration of ethylene oxide was high enough to satisfy the maximal rate of reaction, increasing the ethylene oxide concentration had no effect and the reaction exhibited zero-order characteristics. The minimal concentration of ethylene oxide above which the reaction was zero order increased as the temperature decreased. At 54 C the minimal concentration producing results in agreement with the minimal TCDT curve was 250 mg/ liter. At 40.6 C this value was 440 mg/liter and at 33.4 C, 880 mg/liter. As the concentration of ethylene oxide gas was reduced below the minima for zero order, individual logarithmic time-temperature curves resulted, diverging from the minimal TCDT curve as represented by those for concentrations of 440 and 880 mg/liter. Because they are nonparallel curves, their temperature quotients differ. As calculated from the curves, these are 3.2 for 440 mg/liter, 2.3 for 880 mg/liter, and approach 1.8 as a minimal limit for Qio value. The data from experiments in which mixed responses were found were used to test the sensitivity of the system and to more firmly establish the curves in Fig. 3. Points corresponding to the mixed sets were established on their respective temperature-log time plots. Those that fell within 10 % of the previously established curves were arbitrarily designated "borderline." The number of borderline sets was then compared to the total mixed sets. Three

groups were evident on the basis of this comparison (Table 1). Seven sets were observed with only one of ten tubes showing growth. Although only two of these fell in the borderline category, the remaining five fell to the right or "sterile" side of the curve. The appearance of 1 bead in 25 with a count in the range of 6.15M) to 6.35 logs would indicate that, with 3 beads per tube, one tube in ten would be expected to prove not sterile after exposure to ethylene oxide under minimal conditions for the sterilization of the remainder of the bead population. These sets were therefore recorded as sterile. The second borderline group consists of mixed sets in which two to five tubes of ten showed growth. In this case the number of borderline sets is nearly equal to the total number of sets, indicating that the conditions of the experiments very nearly satisfied the TCDT of the organism. The remaining mixed sets were recorded as not sterile where six or more tubes per set showed growth, and only two sets in ten were in the borderline category. All such points fell to the left of their respective curves. In view of the wide variation and apparent double distribution of

3. Thermochemical death time curves for spores of Bacilluts var. niger dried on glass beads. 0, sterile sets; 0, sets not results of mixed sets.

LITERATURE CITED FRIEDL, J. L., L. F. ORTENZIO, AND L. S. STUART. 1956. The sporicidal activity of ethylene oxide as measured by the A.O.A.C. sporicide test. J. Assoc. Offic. Agr. Chemists 39:480-483.

440 MG /L

4035-

30-

I

25-

20 54 50-

45-

z

40-

a.

X

B

o

880 MG/L

35-

30250

10

20-

ISN-1i 54-

c

N1-

I'l

50I500 MG/L

4540--

3530-

0

-*-

-*

I

25-

2010

10 15

20

30

40

60

80

100

150

200

TIME (MINUTES) FIG.

sutbtilis sterile;*

DISCUSSION

300

A

45-

a:

the spore populations, the transition from conditions allowing recovery of viable spores to conditions producing sterility was surprisingly sharp. In the range of concentrations of ethylene oxide from 22.1 to 884 mg/liter, and of temperatures from 5 to 37 C, the dependence of TCDT on concentration reported by Phillips (1949) is readily apparent. Beyond these limits. in the direction of either increased concentration or increased temperature, Phillips' conclusion that the Ct9o values (concentration X time for 90 % reduction in viable population) are constants for each temperature is not confirmed by our data. This is readily apparent upon examination of Fig. 3. Phillips' data are not sufficiently complete to justify such a conclusion in much of the above range. Although statistical analysis cannot be applied, his data seem to reveal a trend toward high values for Ct9o at high concentrations of ethylene oxide. This effect is predictable in accordance with the present study. Since they are not constant at a given temperature, Ct9o values provide no advantage in representing characteristics of ethylene oxide systems. As Phillips assumed constancy of this value, he based computation of the temperature quotient on the average Ct9o values for the complete range of concentrations, resulting in an approximation of 2.74 for the Qlo of his system. In view of the results reported here, the picture is not so simple. For each concentration of ethylene oxide, there are at least two values for Qlo, one for the temperature range wherein that concentration is limiting to the rate of reaction, and one for the temperature range in which that concentration is greater than the minimum for zero order. In the latter case, the Qlo is independent of concentration, and within the limits of this study, has a value of 1.8. When concentration is limiting to the rate of reaction, the Qlo varies inversely with concentration and is always greater than 1.8. In the light of these results, extrapolation of Phillips' data to higher temperatures and concentrations is invalid and will lead to erroneous figures for sterilization times. The zero-order relationship in a chemical reaction exists where the reacting substance is not limiting to the reaction rate. If this characteristic were to change to first order in a chemical system, such change would be expected as temperature increased. It must be concluded, therefore, that the concentration dependence as temperatures are lowered, appearing in the reaction system involving ethylene oxide and cellular materials, is the result of physical phenomena of adsorption and diffusion and does not reflect a specific chemical reaction per se.

c 5450-

w

[VOL. 10

ERNST AND SHULL

.340

1962]

ETHYLENE OXIDE CONCENTRATION AND TEMPERATURE

KAYE, S. 1949. The sterilizing action of gaseous ethylene oxide. III. The effect of ethylene oxide and related compounds upon bacterial aerosols. Am. J. Hyg. 60:289-295. KAYE, S., AND C. R. PHILLIPS. 1949. The sterilizing action of gaseous ethylene oxide. IV. The effect of moisture. Am. J. Hyg. 50:296-306. PHILLIPS, C. R. 1949. The sterilizing action of gaseous ethylene oxide. II. Sterilization of contaminated objects with ethylene oxide and related compounds: time, concentration and temperature relationships. Am. J. Hyg. 50:280-288. PHILLIPS, C. R. 1961. The sterilizing properties of ethylene oxide, p. 59-75. In The Pharmaceutical Society of Great Britain,

341

Recent developments in the sterilization of surgical materials. The Pharmaceutical Press, London. PHILLIPS, C. R., AND S. KAYE. 1949. The sterilizing action of gaseous ethylene oxide. I. Review. Am. J. Hyg. 50:270-279. REDDISH, G. F. 1957. Methods of testing chemical sterilizers, p. 152-157. In G. F. Reddish, Antiseptics, disinfectants, fungicides and sterilization, 2nd ed. Lea and Febiger, Philadelphia. SMITH, N. R., R. E. GORDON, AND F. E. CLARK. 1952. Aerobic spore-forming bacteria. U. S. Dept. Agr. Monograph No. 16. ZNAMIROWSKI, R., S. McDONALD, AND T. E. Roy. 1960. The efficiency of an ethylene oxide sterilizer in hospital practice. Can. Med. Assoc. J. 83:1004-1006.