Enhancing Service Capabilities by Adding Electron-Beam Irradiator to Gamma Irradiation Facility

Sterigenics had an existing gamma irradiation facility in Jarinu, Brazil, that processed materials and provided irradiation services. In response to increasing demand for irradiation services in Brazil, a decision was made to increase the capacity of the facility by adding a new irradiator. In 2013, a 10-MeV electron beam (e-beam) accelerator was added to complement the gamma facility in Jarinu.

The strategic decision to implement an e-beam accelerator at a facility that already housed a gamma pallet irradiator had the following primary objectives:

1. Diversification of the technology offered to the local market, where Sterigenics already operated two gamma irradiators (at facilities in Cotia and Jarinu).

2. The opportunity for rapidly increasing overall capacity.

3. Adding a complementary technology to the existing gamma pallet irradiator in Jarinu, with the possibility of cross-validation for selected products.

A plan was prepared with targeted products for the cross-validation. The primary factors driving the overall success of cross-validation at the Sterigenics facility in Brazil were the product makeup and the process definition (D_max,acc and D_ster) for the targeted products. (Note: D_max,acc is the maximum acceptable absorbed dose for product, as established by the manufacturer [typically in units of kGy]. D_ster is the minimum required dose for product, as established by the manufacturer [typically in units of kGy]. Ratio of D_max,acc to D_ster is the ratio of the maximum acceptable absorbed dose to the minimum required absorbed dose.)

Existing irradiation processes in the gamma irradiator normally were conducted with the minimum required dose established using dose setting Method 1 or 2 per ANSI/AAMI/ISO 11137-2.¹  On occasion, dose setting was conducted to obtain a sterility assurance level greater than 10⁻⁶ (e.g., 10⁻²). This resulted in minimum dose specifications that were substantially lower than those for sterilization processing, for which the sterilization dose is substantiated using Method VD_max (verification dose maximum).

For most targeted products in the cross-validation, in addition to the low minimum dose requirement, the established maximum acceptable dose was greater than the required minimum by a factor of at least 2.5. Together with the product's packaging specifics, this meant that the cross-validation could result in an e-beam irradiation process capable of obtaining product irradiated within its specification established for gamma irradiation without changing the packaging system.

Figure 1 summarizes the situation at the start of the cross-validation project. The data presented in Figure 1 probably would not be representative of many gamma processing sites, for which a vast majority of processing specifications might be in the range of 25 to 40 kGy or 25 to 45 kGy (ratio of D_max,acc to D_ster of 1.60 to 1.80, approximately).

Even with fairly wide specified ratios of D_max,acc to D_ster, certain products were not cross-validated; these are highlighted in red in Figure 1. In addition, the examples highlighted in yellow required changes to the packaging system^* to be successfully cross-validated. (^*Note: A possible increase in D_max,acc was not pursued for the examples presented.) Additional products that were not considered for cross-validation are not included in Figure 1; however, examples of all situations are included in this article.

A plan was prepared with targeted products for the cross-validation. These primarily were the products highlighted in green in Figure 1.

The starting point for this project was to establish a core, multidisciplinary group within the company to partner with selected manufacturers for the cross-validation. With e-beam irradiation being a new service offering in Brazil, an understanding of the new technology, considering similarities and differences with respect to gamma irradiation, were discussed and explained to the manufacturers. The manufacturers primarily were interested in the increased service offering from e-beam and the possibility of having an alternative technology as a backup. The addition of the e-beam technology also offered additional operational flexibility without compromising the continuity of the process services already performed at the facility.

In Brazil, healthcare products go through regulatory registration approval, during which the technology or technologies used for sterilization need to be defined. The addition of the e-beam technology for sterilization of a medical device requires a revision to the device's registration. This generally takes about 12 months for medical devices that are being cross-validated; however, it could take longer in some cases. For labware and pharmaceutical packaging, the approvals generally take three to six months. For other product categories (e.g., agricultural products), no regulatory approvals are required for cross-validation. When required, regulatory approvals are managed by the manufacturer of the product.

To ensure compliance with the regulatory requirements and applicable standards for medical device sterilization, the sterilization dose for a medical device needs to be established and its effectiveness demonstrated when product is irradiated using an irradiation source different from that for which the sterilization dose was established. ISO 11137-1 states that such evidence is provided through a successful audit of the sterilization dose, with the verification dose irradiation occurring using the radiation source to which transfer is considered.²  For all medical devices, labware, and pharmaceutical packaging in the cross-validation, a successful verification dose experiment was performed with the verification dose irradiation performed at the new e-beam irradiator. The verification dose and sterilization dose, or more generally the minimum required dose, were cross-validated in all cases.

In parallel with the demonstration of continued effectiveness of the minimum dose, the performance attributes of the different products and their composing materials with respect to the e-beam irradiation process were evaluated. The strategy aimed to, at first, individually evaluate the various products through a qualitative evaluation and a visual comparison with gamma-irradiated samples. Further pursuit or rejection of the potential cross-validation was determined based on the qualitative evaluation by the manufacturer.

For the vast majority, the qualitative behavior of the products and their composing materials were very similar to those observed in gamma irradiation. Some positive variations were seen in terms of behavior for e-beam (e.g., some polypropylene exhibited less discoloration in e-beam versus gamma) and negative for other materials (e.g., some rigid polyvinyl chloride [PVC] products exhibited more discoloration in e-beam versus gamma). These differences may or may not be noteworthy factors in the ability to cross-validate the product, depending on the product application. In the case of the product containing the rigid PVC that was exhibiting more pronounced discoloration after e-beam processing, this observation ended the attempt for cross-validation.

After a high-level screening, as described above, select products became part of a formal program to transfer the maximum acceptable dose established for gamma irradiation to e-beam. Product samples were irradiated in an as-uniform-as-practicable manner and at a selected dose corresponding to that established as the maximum acceptable dose for gamma irradiation. Irradiated product was returned to the manufacturer for testing of performance attributes. Under the controlled conditions of the irradiations, the maximum acceptable dose could be transferred to e-beam for all examples in this article, with the exception of the example of rigid PVC noted earlier.

Performance qualification (PQ) dose mapping was performed either concurrently or shortly after completion of transfer of the minimum required dose and transfer of the maximum acceptable dose.

In addition, for product validation purposes, some manufacturers performed irradiation at the conditions established for routine processing. The agricultural growth media discussed in this article were examples for which such an approach was taken.

The total time from initiating the cross-validation project to submission for approval for processing in e-beam ranged from six months for labware products and pharmaceutical packaging to a maximum of 18 months for certain medical devices. Agricultural and food packaging products could be cross-validated in a shorter time frame, if successful.

For medical device products validated following a change in product packaging (Figure 1), the cross-validation followed the steps described above but took longer because of the packaging modifications.

In general, medical devices and labware represent a category of relatively low-bulk-density products in their presentation (bulk density typically <0.15 g/cm³). However, for e-beam, the possible and inconsistent overlap of materials at points in the final transport packaging, characterizing a heterogeneous product density distribution, must be considered in substantially more detail compared with gamma. The overlapping of layers of a polymeric component, for example, can generate a high localized density and hinder the local penetration of radiation by e-beam (see example Medical Device 4 below).

Understanding, predicting, and measuring dose distribution was one of the biggest challenges in cross-validation between the technologies. For e-beam, PQ dose mapping generally is specific to and defined according to individual products. Minimizing variability among shippers by the manufacturer is essential to keeping the dose distribution and magnitude within an acceptable degree of variability and to reducing the amount of testing required for PQ dose mapping.

Certain products have a large and systematic overlap of materials in the shipping packaging. The e-beam dose distribution for these products can be largely incompatible with the product dose specifications established by the manufacturer. The food and pharmaceutical packaging products for which cross-validation was not further pursued (Figure 1) are examples of this.

Cross-validation from gamma to e-beam sterilization for this personal protective equipment (PPE) medical device was the first successfully completed for this category of products. The project went through the stages of transfer of the maximum acceptable dose and auditing of the sterilization dose, with the PQ dose map carried out in parallel. Successful cross-validation to e-beam from gamma resulted from several factors:

1. Relatively large degree of homogeneity and consistency of the product in the shipper

2. Possibility to orient the shipper in a way that resulted in a relatively small penetration depth of 28 cm for the e-beam, given the relatively high bulk density of 0.20 g/cm³

3. A high specified ratio of D_max,acc to D_ster of 5.00

4. A relatively low maximum acceptable dose; therefore, possible effects from adiabatic heating in e-beam processing were not found to be a practical concern

The fact that this product had a very wide dose specification was critical to the success of the cross-validation. The product was irradiated through the narrowest available depth of 28 cm in e-beam, and the resultant average dose uniformity ratio (DUR; the ratio of the maximum to minimum absorbed dose, as measured in the irradiation container) using double-sided irradiation was 2.65. Had the specified ratio of D_max,acc to D_ster been tighter (e.g., a D_ster of 25 kGy and D_max,acc of 50 kGy results in a ratio of 2.00), the product could either not have been processed in e-beam or would have required considerable changes to the product packaging and/or the maximum acceptable dose.

This product was easily converted to e-beam from gamma because of several factors:

1. Relative homogeneity of the product in the shipper boxes

2. A relatively low bulk density of 0.08 g/cm³

3. A relatively high ratio of 2.86 between D_max,acc and D_ster

The fact that this product was low bulk density, was relatively uniform, and had a relatively wide specified ratio of D_max,acc to D_ster of 2.86 was critical to the success of this cross-validation. The resultant average DUR of 1.52 in e-beam was acceptable for this product, but when process variability and uncertainty were incorporated, the process would not have been acceptable for irradiating product with a ratio of D_max,acc to D_ster of 1.60.

This product was cross-validated in e-beam from gamma. However, a packaging change was required in order to achieve an acceptable dose distribution. The specified ratio of D_max,acc to D_ster of 2.45 was not possible to achieve in e-beam, as a result of the relatively high bulk density of 0.17 g/cm³ and the specifics of the packaging method.

Initial PQ dose mapping of this product took place by irradiating the product through the narrowest penetration depth afforded by the product packaging (28 cm). The resultant average DUR of 2.40 was deemed to be not sufficiently low for rendering a process capable of obtaining product irradiated within its defined specification, despite being below the specified ratio of D_max,acc to D_ster of 2.45. Instead, a packaging change was performed by the manufacturer to allow irradiation through a penetration depth of 21 cm. The PQ dose mapping resulted in an average DUR of 1.80.

The fact that this product had a specified ratio of D_max,acc to D_ster of 2.45 was not enough to readily overcome the challenge of cross-validation in e-beam. This wide range was more than sufficient for gamma; however, a packaging change was required to achieve a process in e-beam deemed sufficiently capable for irradiating product within specification. Developing, validating, approving, and implementing the packaging change resulted in a substantial delay in the cross-validation compared with the examples discussed previously (i.e., medical devices 1 and 2).

Possible variability in the configuration of the product inside a shipping box should be assessed in e-beam processing.

Medical device 4 consisted of bloodlines packaged in a coiled manner. Inherently, this results in an area of large density in which the bloodline is coiled up and the presence of a central void area. For gamma irradiation, the orientation and relative position of the bloodlines in the shipper were not controlled to the degree required for an e-beam process to allow product irradiation within the value of 2.67 established for the ratio between D_max,acc and D_ster.

The manufacturer had to ensure a more consistent positioning of the bloodlines with respect to one another. Finally, a method was established where the bloodlines were stacked side by side in a shipper and in a manner where the shipper could be irradiated with the e-beam having to penetrate only through a single bloodline. This configuration rendered on average a DUR of 2.25, which was sufficient for routine irradiation within specification.

In general for labware, the resultant DUR in e-beam was substantially greater than that achieved for the same product in gamma radiation. This did not represent limitations for the vast majority of products cross-validated, as the minimum dose required for most labware was low and the specified ratio between D_max,acc and D_ster was quite wide. Completion of the cross-validation for labware was considerably faster than that for medical devices because of the different approvals needed.

Labware product 1 was empty petri dishes, with a specified ratio of D_max,acc to D_ster of 3.20, bulk density of 0.10 g/cm³, and target penetration depth of 25 cm. The achieved average DUR in e-beam was 2.04, and the process was deemed capable of irradiating product within its specification.

The fact that this product was relatively low bulk density, was relatively uniform, and had a relatively wide specified ratio of D_max,acc to D_ster of 3.20 was critical to the success of the cross-validation.

Labware product 2 was empty flasks, with a specified ratio of D_max,acc to D_ster of 3.40, bulk density of 0.03 g/cm³, and target penetration depth of 49 cm. The achieved average DUR in e-beam was 1.70, and the process was successfully cross-validated.

The fact that this product was very low bulk density, was relatively uniform, and had a relatively wide specified ratio of D_max,acc to D_ster of 3.40 was critical to the success of the cross-validation. The combination of all these factors resulted in a good average DUR of 1.70 despite the large target penetration depth of 49 cm.

This product was easily converted to e-beam from gamma because of several factors:

1. Relative homogeneity of the product in the shipper boxes

2. A very low bulk density of 0.04 g/cm³

3. Relatively narrow target penetration depth of 33 cm

The fact that this product was very low bulk density and was relatively uniform resulted in a process deemed capable of achieving dose within its specification, even though the specified ratio of D_max,acc to D_ster was 2.25. The resultant average DUR in e-beam was 1.73.

The processing of higher-density products by e-beam can be difficult because of the limited penetration of e-beam radiation through the product. One of several interesting experiences in the cross-validation of gamma processes to e-beam was related to a product used in the agriculture segment.

Because of the product's high bulk density (0.44 g/cm³), it might be thought that processing the material using e-beam would not be possible. This particular product had a packaging system that provided a narrow penetration depth of only 15 cm and a relatively wide specified ratio of D_max,acc to D_ster of 3.33. Using double-sided irradiation, on average a DUR of 2.00 was achieved for this product, and at first, it was deemed successfully cross-validated to e-beam.

However, during testing of the product at routine processing conditions, it was observed that in the e-beam irradiation process, the product substantially absorbed and retained heat due to the high density and self-insulating properties. The temperature increase was assessed to be sufficient to possibly compromise the primary and secondary polymeric packaging.

To reduce the temperature buildup, the e-beam process finally was carried out in two stages, with an interruption of 24 hours between each stage. The interruption period allows for the product's temperature to return to ambient while not compromising the effectiveness of the minimum dose that was established.

This example demonstrates that it could be good practice during cross-validation, or validation in general, to not solely consider irradiating in an as-uniform-as-practicable manner but also to supplement those studies with irradiations performed under conditions mimicking routine irradiation.

Rubber stoppers for closures of pharmaceutical packaging typically are provided to end-users in bulk quantity. For the specific example in this article, the product as presented for gamma irradiation had a high bulk density of 0.37 g/cm³. The specified ratio of D_max,acc to D_ster was 2.50.

This product could not be cross-validated in e-beam because the PQ dose mapping measured zero dose near the center of the product within the sterile barrier, even though the narrowest penetration depth for the packaging of 22 cm was used. The rubber stopper's density for the given penetration depth was too high and the e-beam was unable to effectively penetrate the entire product. The packaging changes that would allow to respect the dose specification were too stringent in nature for the manufacturer and the end-user to accept; therefore, the product was not able to be cross-validated in e-beam.

Food packaging typically has a relatively high bulk density, and this product had a bulk density of 0.24 g/cm³. The specified ratio of D_max,acc to D_ster was 3.00. The average DUR achieved in e-beam was 3.00, with the product irradiated through the smallest penetration depth of 25 cm. This process was not capable for rendering product irradiated within its specification once appropriate processing buffer was applied, and product could not be repackaged. Therefore, the product was not able to be cross-validated in e-beam.

In another example of an agricultural-type product, the cross-validation to e-beam was not economically viable. In this instance, the heat absorbed and retained by the product caused damage to the primary packaging of the product, despite the introduction of a cool-down period and an attempt to process the product at lower power levels of the accelerator. Processing exclusively using gamma radiation was continued.

The processing speed generally is considered to be a favorable characteristic in e-beam processing compared with gamma irradiation. This effectively is an advantage for many products; however, for processes of higher-density or temperature-sensitive materials, the generation of heat in an adiabatic manner and the possible inability of the product to dissipate the heat in a timely manner can cause product and/or packaging damage. This can be overcome by performing the irradiation at lower power levels of the accelerator or by introducing a fragmented irradiation process that possibly includes a cool-down period. These actions will increase the total processing time of the product.

• Products containing powders or liquids generally were not considered for cross-validation in e-beam because the powder or liquid could flow in the packaging system during product movement. When inverting for a second pass through the radiation field, which is general practice for improving the uniformity of irradiation at the installed irradiator, this movement leads to a complication of determining the dose effectively received by the powder or liquid. Irradiation from a single side only therefore could be preferred at the facility for such product. Further complication could be introduced due to the large localized density and/or mass of the powder or liquid and the fact that the installed dosimetry system was not validated for use in a liquid environment. These factors would need to be considered before attempting to validate product with these materials.

• Products with expected dose gradients over a distance that is too short for the installed dosimetry system to provide sufficient spatial resolution (e.g., products containing metal components).

• More generally compared with the specific situations for powders, liquids, and metals described above, products for which it was not physically possible to place a dosimeter at all positions for which the dose needed to be determined.

Cross-validation of products irradiated in gamma to e-beam irradiation includes transferring, and sometimes reestablishing, minimum required and maximum acceptable dose, as well as successful PQ dose mapping.

The Sterigenics facility in Jarinu, Brazil, was successful in cross-validating a large number of products from gamma to e-beam, thereby providing complementary technology offerings and optimization of resources. This success was largely attributable to the favorable product dose specifications and packaging specifics for the primary products intended for cross-validation. These favorable specifications were due to a number of factors, including the method of minimum dose validation, the use of sterility assurance level levels other than 10⁻⁶, and other factors not specifically covered in this article. Without the presence of these factors, completing the cross-validation would not have been possible in the projected time frame.

Both gamma and e-beam sterilization technologies continue to be used extensively at the Sterigenics facility, in order to successfully process the substantial diversity of product types and dose specifications.