A Discriminating Method for Measuring the Microbial Barrier Performance of Medical Packaging Papers

Microbial barrier performance is a critical factor in the selection of packaging materials for use with sterile medical devices. Previous reports of our research in this area have stressed the importance of overall flow rate in determining the extent of penetration of medical webs by air-dispersed bacterial spores.^1­3 In two earlier studies that used overall flow rates between 10^-3 and 10^-1 dm³ min^-1 cm^-2 (liter per minute per square centimeter), uncoated grades of medical Tyvek (designated 1059B and 1073B by DuPont) showed maximal penetration values that were grade specific and independent of the concentration of the air-dispersed spores.^4,5 Over similar flow rates, the extent of penetration of medical packaging papers generally decreased with increasing flow rate and no maximal values were evident.

At the time of those studies, practical considerations limited the range of flow rates over which penetration could be measured, but it was believed that if measurements could be conducted at sufficiently low rates, a maximal value of penetration (P_max) characteristic of the paper under test would be revealed.² It is our view that such a value would define the barrier performance of the material toward airborne bacterial spores; in essence, the P_max value would quantify the efficiency of the material at stopping the passage of such organisms.

A new test apparatus has now been developed to measure the microbial penetration of paper packaging materials at rates of flow low enough to detect a P_max value. A description of that system is presented in this article, along with the results of qualification studies carried out on the apparatus. Also presented are measurements of P_max made on several commercially available papers that are used for medical packages sterilized by exposure to ethylene oxide, irradiation, or steam.

Test Organism. As in previous studies, endospores of Bacillus subtilis var. niger were selected as the test organism; they are nonpathogenic, exist as discrete cells, produce readily measurable colonies, and are robust and thus able to withstand the physical stresses associated with the test procedures. Furthermore, the spores, which are elliptical in shape and have aerodynamic characteristics similar to those of a 0.9-µm-diam spherical particle, are highly stable when individually dispersed in air.

The test spores were produced as a surface growth on tryptone glucose yeast agar (Oxoid Ltd., Basingstoke, England) after 7 days incubation at 37°C. They were then collected in water, washed three times in water, and stored at 4°C as a suspension with an approximate concentration of 5 X 10⁸ spores cm^-3. The resulting batch of spores was used over an experimental period of between 3 and 6 months, during which time no significant loss of viability was observed.

Barrier-Test Apparatus. The central feature of the test apparatus used for our previous studies was a static reservoir containing air-dispersed spores held under stirred-settling conditions.² Measurements of spore penetration were conducted at overall flow rates down to 10^-3 dm³ min^-1 cm^-2, but lower test flow rates were precluded because of a gravitational loss of spores from the flowing dispersion. Such gravitational losses had been eliminated by other researchers who used a containment chamber that rotates around a horizontal axis.⁶ Our new test apparatus built upon that concept to provide a system for measuring spore penetration at ultralow flow rates.

Figure 1 is a sectional diagram of the apparatus's rotating chamber, in which a dispersion of spores in air is produced, contained, and presented as a challenge to the test sample. Consisting of a sealed cylindrical drum constructed of 2.5-mm-thick sheet aluminum, the chamber measures 60 cm long and 60 cm in diameter. It is supported on a fixed benchtop cradle, and two bearings coupled to extensions at the ends of the drum allow it to rotate freely around the horizontal axis.

During the tests reported on here, the chamber was rotated via a drive motor, suitably geared and controllable, at a fixed number of revolutions per minute ranging from 0.3 to 5.2. One end of the chamber was sealed, using two grooved O-rings, to a stator that formed part of the cradle. Four access ports--designated N, E, S, and H--were located in the stator. Two of these ports were used to charge the chamber: Port N was connected directly to the outlet of a modified Collison nebulizer⁷ and port E to a high-efficiency filter (sodium flame test­ rated at >99.9999% efficient). The latter port acted as a vent to atmosphere, which allowed equilibration of pressure during aerosolization and sampling. A humidity and temperature probe (HMP 31; Vaisala, Helsinki, Finland) was permanently mounted in the chamber through port H, and a permanent linear brass sampling tube (ID 0.8 cm) extended 12 cm into the chamber through port S. When not in use, the sampling tube was capped at its external end. The other end of the chamber, opposite the stator, rotated with the drum body, and it was through this end that the spore challenge was presented, via a 10-cm-diam opening, to the paper samples being evaluated.

Figure 2 is a schematic of the test assembly that is used to hold the paper sample vertically in the flow of spore dispersion, to collect the penetrating spores, and to set the overall flow rate. The circular test sample is held between two grooved neoprene O-rings. Other assembly components include a gelatin filter with an average pore size of 3 µm (Type SM 12602 Alk; Sartorius AG, Göttingen, Germany), and a calibrated critical-flow orifice (CFO) for setting the overall flow rate. Both the filter and the CFO are also compressed against neoprene O-rings to ensure leak-free seals.

The test assembly can be attached to or removed from the chamber (at location T in Figure 1) while the chamber is rotating. The integrity of the contained spore dispersion during connection and disconnection is maintained via operation of a gate valve.

During testing, samples with an area of 78.5 cm² were exposed to the flow of air- dispersed spores. By use of a series of CFOs, the volumetric flow rate through the assembly was set at a value between 8.7 X 10^-3 and 1.0 X 10¹ dm³ min^-1, corresponding to overall flow rates of 1.1 X 10^-4 to 1.3 X 10^-1 dm³ min^-1 cm^-2 paper sample (the overall flow rate equals the volumetric flow rate [liters per minute] divided by the test sample area [78.5 cm²]).

Test Papers. Twelve uncoated papers (designated A through L) were supplied as 35-cm-wide rolls, and 12-cm-diam circular samples were cut from random locations along each roll. The weight of the 12 papers ranged from 45 to 100 g m^-2 and their air permeability extended from 46 to 650 Bendtsen. Prior to barrier testing, the paper samples were conditioned by exposure to air at 23°±1°C, with relative humidity at 50±2%, for a period exceeding 4 hours, according to procedures in British Standard 3431 (1973).

The assessments of the microbial barrier properties of the 12 packaging papers followed a sequence of procedural steps.

Dispersion Setup. The first steps involved the conditioning of the chamber, the dispersal of spores in an aerosol, and the introduction of this dispersion into the chamber. With the chamber static, the gate valve was closed and filtered air, dried by compression, was introduced into the chamber via port S at a flow rate of 10 dm³ min^-1 and vented to atmosphere through port E. This operation was continued until the relative humidity in the chamber reached a level of about 40%. (Preconditioning of the chamber environment in this manner ensured that the relative humidity of the contained air was set at 50% after the chamber was charged with air-dispersed spores via nebulization of the aqueous suspension.)

In the next step, with the chamber rotating at 3 rpm, a spore suspension of appropriate concentration was nebulized into the chamber at 2 bar for 20 minutes. Typically, attainment of a dispersion with a concentration of 10⁶ spores dm^-3 required nebulization of a suspension containing 5 X 10⁸ spores cm^-3, and a dispersion concentration of 10³ spores dm^-3 required nebulization of a suspension of 5 X 10⁵ spores cm^-3.

Determination of Challenge Concentration. For a given working day, the concentration of dispersed spores in the chamber was calculated as the mean of two estimates, one made at the start of the workday and the other at the end. Each estimate was obtained by drawing a 1-dm³ sample of spore dispersion through a gelatin filter at a rate of 0.5 dm³ min^-1. The filter was mounted in an in-line 50-mm filter holder (Type SM 165 98; Sartorius AG, Göttingen, Germany) that was connected directly to the outer end of the brass sampling tube extending into the chamber through port S. After sampling, the gelatin filter was recovered and dissolved in a known volume of sterile aqueous solution of 0.25% sodium alginate to release the spores to suspension. Viable spores collected from the filter were then enumerated using a standard surface plate count method.

Spore Penetration Measurement. For each test run, a paper sample, a gelatin filter, and the appropriate CFO were mounted in the test assembly. Then, with the chamber rotating, the test assembly was attached to the face of the gate valve mounting and a connection made, via the rotating joint, to a vacuum source. At that point, the gate valve was opened and vacuum applied for a period sufficient to allow 1 dm³ of spore dispersion to be drawn through the test sample. The time required to achieve this volume varies depending upon the flow rate employed, which, in turn, is dependent upon the choice of CFO. In this study, the test period ranged from 0.1 minute to 2 hours.

At the end of the test period, the vacuum was disconnected, the gate valve closed, and the test assembly disassembled. Viable spores collected on the gelatin filter were then released to suspension and enumerated as described above. An estimate of the percentage (P) of spores penetrating the test sample when challenged at a defined flow rate was obtained using the following equation:

P = number of penetrating spores dm^-3 /number of challenge spores dm^-3 X 100

Estimates of percentage penetration were carried out at seven different challenge flow rates ranging between 1.1 X 10^-4 and 1.3 X 10^-1 dm³ min^-1 cm^-, each estimate being made on a separate sample.

Prior to the comparative study of 12 packaging papers, several qualification tests were carried out on the new apparatus.

Characterization of the Contained Spore Dispersion. The decay characteristics of the spores held in dispersion within the rotating chamber were determined at seven different rotational speeds ranging from 0.3 to 5.2 rpm. At each speed, the chamber was charged with air-dispersed spores (at a nominal concentration of around 10⁵ spores dm^-3), and then, following cessation of nebulization, the dispersion concentration was determined a number of times within a 50-hour period. A log plot of spore concentration against time gave a linear decay curve from which an estimate was made of the dispersion's half-life (i.e., the time, in hours, in which the dispersion underwent a twofold reduction in spore concentration).

Figure 3 shows an arithmetic-scale plot of dispersion half-life versus rotational speed. At a speed of 0.3 rpm (the lowest examined), the dispersion half-life was around 20 hours; progressive increases in the rotational speed resulted in corresponding increases in dispersion half-life, so that a value of about 55 hours was seen at 2 rpm. However, further increases in rotational speed yielded a progressive decrease in half-life. This behavior is in accord with that described elsewhere for particles dispersed within a rotating chamber.⁸ These results demonstrate that, within the rotational speed range of 1­3 rpm, spore dispersions within the chamber of the new apparatus exhibit a high level of stability, with half-lives exceeding 48 hours.

Additional half-life measurements using spores dispersed at various concentrations between 10² and 10⁶ spores dm^-3 and a fixed chamber rotation speed of 3 rpm yielded results that did not differ significantly (p = 0.05). This lack of dependence of dispersion decay on spore concentration is again in keeping with other research.⁸

Their proven high stability level allowed us to use spore dispersions in challenge studies throughout a 6­8-hour workday, during which time the dispersion concentration remained effectively constant. Table I lists estimates of dispersion concentration made at the start and end of 10 working days. Although the dispersion concentrations were generally lower at the end of the day than at the start, the magnitude of the differences was small, with an average reduction factor of 1.09 (i.e., an 8% reduction). Based on these data, the practice was established of determining dispersion concentration for a given day from the mean of two spore concentration estimates, one made at the start of the working day and the second at its conclusion.

Spore Collection Efficiency at Different Flow Rates. When performing measurements of spore penetration at ultralow flow rates, a key requirement is the elimination of the effects of gravity on the spores in dispersion. Therefore, the qualification testing of the rotating chamber apparatus included experiments aimed at validating the absence of gravitational spore loss over the operable range of flow rates.

Following the procedural steps described above, a spore dispersion with a concentration of ~10⁵ spores dm^-3 was generated within the rotating chamber (3 rpm) and daily determinations were made of the actual concentration within the chamber. At each of a number of different flow rates, a fixed volume (1 dm³) of spore dispersion was drawn through the test assembly, which was attached to the chamber without a paper sample. For each flow rate an estimate was made of the number of spores collected on the assembly's gelatin filter, and this result was then expressed as a percentage of the corresponding spore dispersion concentration, providing a measure of spore collection efficiency.

Figure 4 is a plot of spore collection efficiency versus volumetric flow rate. It is clear from the figure that, over the wide test range of volumetric flow rates, collection efficiency values were grouped around 100% (the dashed line). Moreover, the correlation coefficient derived from the data did not differ significantly from zero (p=0.05); in other words, collection efficiency is independent of flow rate. This response demonstrates both the robustness of the test system and its suitability to undertake controlled challenge studies over a wide range of flow rates extending to ultralow flows with no apparent loss of airborne spores due to gravity.

Test Variability. To assess the variability that can be attributable to test procedures, as opposed to that associated with paper samples, six separate samples of a typical medical packaging paper (paper C) were challenged at a fixed flow rate (6.4 X 10^-3 dm³ min^-1 cm^-2) on five different occasions and spore penetration percentages were determined as described earlier.

Table II lists the derived percent penetration values and Table III provides the results of a two-way analysis of variance performed on those data. The analysis revealed the following:

One paper type was chosen to illustrate our general findings for medical packaging challenged with air-dispersed spores flowing at rates extending to ultralow levels. Figure 5 shows a logarithmic-scale plot of spore penetration percentage versus flow rate for paper C. At the lowest flow rate, of approximately 10^-4 dm³ min^-1 cm^-2, penetration is about 0.6%, and this value initially increases progressively as the flow rate increases, reaching a level of about 1% at a flow rate of around 8 X 10^-4 dm³ min^-1 cm^-2. At that point, penetration begins to decrease with increasing flow rate, a complete change in behavior. Thus, it can be seen that there is a particular flow rate where penetration is maximal (P_max).

Figure 6 shows plots of penetration versus flow rate for all 12 medical papers evaluated (A through L). These plots reveal that, for each paper type, there is a characteristic relationship between penetration and flow rate resulting in the occurrence of a P_max value specific to the paper type. However, it is also clear that the overall shape of the curve relating these parameters differs from paper to paper. For example, in response to a 100-fold change in flow rate, penetration varies over a 17-fold range for paper I, compared with a less than 2-fold range for paper B.

Table IV lists the observed values of P_max and corresponding flow rates for the 12 papers. These P_max values range from 0.15 to 55% (for papers L and D, respectively), which represents a 360-fold range in values. The flow rate at which P_max falls also differs from paper to paper, ranging from 2.1 X 10^-4 to 3.9 X 10^-3 dm³ min^-1 cm^-2 (for papers F and A, respectively), which represents a 16-fold range in flow rate.

As discussed above, the experimental approach taken to measure the microbial barrier properties of medical papers was to challenge separate 10-cm-diam samples of paper with air-dispersed spores flowing at different rates extending to ultralow levels. At the lowest flow rate examined (10^-4 dm³ min^-1 cm^-2), a test duration of 2 hours was required to conduct a single measurement of spore penetration. Clearly, such prolonged test durations necessitate the use of a spore dispersion that is highly stable. This level of stability was achieved by containing the spore dispersions in a rotating chamber. Under optimal conditions of rotation (~3 rpm), spore dispersions at different concentrations, which extended over a 10,000-fold concentration range, all had half-lives exceeding 48 hours, so that no significant decrease in spore dispersion concentration was detected over a 6­8-hour working day.

The high stability level of spore dispersions held within the rotating chamber was confirmed by tests showing that the spore collection efficiency of the barrier-test apparatus is independent of flow rate. Thus, measurements of spore penetration conducted on individual paper samples are proper measures of the barrier properties of the samples under specified test conditions.

Furthermore, data generated to examine the variability of the test procedure clearly demonstrate that this variability is less than that associated with the sample-to-sample variability of a typical medical paper. (Such a low procedural variability is another necessary prerequisite for effective discrimination between medical papers possessing different barrier properties.)

The test data for the 12 papers evaluated using the new apparatus confirmed previous findings that porous medical packaging materials are not absolute barriers to air-dispersed spores and that the extent of penetration is strongly dependent upon flow rate. At very low flow rates, penetration is directly related to flow rate, whereas at higher flow rates it is inversely related to flow rate. Consequently, there is a flow rate at which penetration is maximal (and spore capture is minimal).^2,4

As explained in a previous article concerned with medical grades of Tyvek, the occurrence of P_max can be understood if the paper sample is regarded as a depth filter opposing the flow of air-dispersed spores.⁴

Over the low-flow-rate domain, Brownian motion of spores and electrostatic attraction between spores and fibers cause spores to leave the airflow and adhere to fiber surfaces. The extent of spore capture by these means is inversely related to flow rate, so that the extent of penetration increases with increasing flow rate.⁹

Over the high-flow-rate domain, however, spores acquire high inertial forces, which predominate and cause the spores to leave the airflow and impact upon, and in turn adhere to, fiber surfaces. Inertia is a direct function of flow rate and thus, over this domain, penetration decreases with increasing flow rate.¹⁰

Given these two opposing behaviors, P_max occurs at local-flow conditions within the paper structure where the cause of spore capture switches from predominantly Brownian motion and electrostatic attraction to capture via inertial forces. The P_max value is a meaningful measure of the barrier performance of the paper as determined by the sheet structure, which enables researchers to quantify the microbial barrier performance of the sheet structure under defined conditions of minimal spore capture.

The varying P_max values observed in this study indicated that commercial medical packaging papers have markedly different barrier properties, which must reflect differences in critical variables that affect sheet structure. However, commercially available papers are not appropriate test materials for seeking a full understanding of the relationships between sheet structure and microbial barrier performance. Generally, the specifications of such papers are not well-defined and various papers are fabricated from different source materials and under different processing conditions; they also have in or on them many different chemicals, such as sizing agents, fillers, and adjuvants. Precisely specified papers, especially made under controlled conditions, are needed to delineate the relationship between structure and microbial barrier performance.

Similarly, the challenge flow rate at which P_max occurs is different for different paper types. Theoretically, the location of P_max is determined solely by the test variables associated with the spore dispersion, particle mass, and fluid (air) viscosity and density. Because these variables were held constant for our study, the expectation was that P_max would fall at common local-flow conditions within the paper structure. In practice, however, local-flow conditions for a given overall flow rate are influenced by paper structure, and, since the latter differs markedly from one commercial packaging paper to another, the locations of P_max on the flow rate axis varied for each paper. Thus, in our view, it is inappropriate to use a single fixed overall challenge flow rate to measure the barrier performance of commercial packaging papers.

In summary, the investigation described above confirmed the following assumptions:

FLOW RATE UNITS

Following scientific practice in Europe, the authors of this article have quoted flow rates in terms of face velocities (overall flow rates) expressed as dm³ min^­1 cm^­2 (liters per minute per square centimeter). These measurement units allow volumetric flow rates (liters per minute) to be normalized for sample size, thereby enabling researchers to compare the outcomes of different barrier test methods.

Note that different face velocities are produced when a given volumetric flow passes through a small sample area (such as the one used in the method outlined in ASTM F 1608-95) as opposed to a large sample area (such as that used in the method described in this article).

Alan Tallentire is a professor of pharmacy at the University of Manchester, England. Colin S. Sinclair is a senior research associate in the Department of Pharmacy, University of Manchester. The authors' research was funded in part by Rexam Medical Packaging (Vernon Hills, IL).

1. Sinclair CS, and Tallentire A, "Penetration of Medical Grade Papers by Airborne Bacterial Spores," J Pharm Pharmacol, 38:77P, 1981.

2. Sinclair CS, and Tallentire A, "Microbiological Evaluation of Papers--The Influence of Dispersion Flow Rate on Penetration," Partic Microb Cont, 1(1):52­57, 1982.

3. Tallentire A, and Sinclair CS, "Sterility Maintenance--Porous Packaging Materials," in Sterilization of Medical Products, vol III, Harris LE, and Skopek A (eds), North Ryde, New South Wales, Australia, Johnson & Johnson, pp 230­246, 1985.

4. Tallentire A, and Sinclair CS, "Microbiological Barrier Properties of Uncoated and Coated Spunbonded Polyolefin (Tyvek)," Med Dev Diag Indust, 7(10):50­54, 1985.

5. Sinclair CS, and Tallentire A, "The Influence of Dispersion Concentration on Microbiological Barrier Performance of Porous Packaging Materials," Med Dev Diag Indust, 8(8):34­37, 1986.

6. Goldberg L, Watkins HMS, Boerke EE, et al., "The Use of a Rotating Drum for the Study of Aerosols over Extended Periods of Time," Am J Hyg, 68:85­93, 1958.

7. May KR, "The Collison Nebulizer: Description, Performance, and Applications," J Aerosol Sci, 4:235­243, 1973.

8. Dimmick RL, and Akers AB, An Introduction to Experimental Aerobiology, New York, Wiley Interscience, 1969.

9. Davies CN, "Filtration of Aerosols," J Aerosol Sci, 14:147­161, 1983.

10. Stenhouse JIT, Harrop JA, and Freshwater DC, "The Mechanisms of Particle Capture in Gas Filters," J Aerosol Sci, 1:41­52, 1970.