Introduction
Single use systems have become an integral part of biopharmaceutical drug manufacturing. The low capital investment and the flexibility of storage capacity appeals to manufacturers because costs are incurred when a product is being produced. Blast frozen bags containing valuable drug product can be efficiently stored in controlled cold room spaces. The bags take up roughly the same space as the product itself. An alternative, multi-use stainless steel cryo vessels, are less efficient for storage and are expensive to validate, clean, and maintain [1].
The Bayer Healthcare site in Berkeley, California is well into its second decade of developing manufacturing processes that take advantage of single use technologies. Bayer pioneered the use of frozen storage bags for product intermediates. Today, bags containing protein product intermediates are routinely blast frozen in plastic bags prior to storage at -30C in a cold storage warehouse. Typically, the bags are handled properly and the process follows as planned. However, frozen bags are slippery, and can be dropped. At low temperatures, the bag integrity may be breached, which results in product loss.
The purpose of this study was to evaluate commercially available bioprocess storage bags to determine if any were superior at protecting the frozen contents. The bags were subjected to simulated filling, freezing, worst case transport, and warehouse handling to determine which were best at resisting damage.
Background
Currently, high value product is stored in blast-frozen poly EVA (ethylene vinyl acetate) bags (Figure 1). EVA is very strong at room temperature, but can be easily damaged at the lower temperatures required for product storage (Figure 2). Often, the loss of a single bag can represent a significant monetary loss.
Damage to bags does not occur until after the product is filled and frozen. Typically, the damage is not observed until the product is thawed, which has a negative impact on three aspects of production: real-time difficulties with production planning, storage of worthless product as inventory, and monetary loss through the damage of product.
We have implemented several process improvements and procedural controls to protect bags and improve handling. However, using a bag better able to withstand worst-case handling would be a superior solution to prevent product losses. Potential replacement bags must be able to resist damage from mishandling at a wide range of temperatures.
Several issues must be considered before changing to a new storage bag. Namely, the new bag must be compatible with the current product, for example, the product must not adsorb to the bag. Product stability must be validated for a different plastic film, and the bag material must be safe to use for biopharmaceuticals. Other factors must also be considered, including bag cost and supply chain reliability [2]. Maintaining the current configuration of the bag minimizes the need for process re-validation.
Materials and Methods
The purpose of evaluating disposable bags is to select the product that will perform best under actual production conditions. The most important performance factor considered during the evaluation is leaking during the freeze-storage-thaw cycle. The experimental designs were intended to mimic as much as possible the personnel handling, storage, and transportation of the frozen material in the storage bags.
Materials
We asked multiple companies to submit bags comparable in volume to the bag currently in use. The scope of this project did not include any process redesign. Currently, the bags we use are stored at -30C. The goal was to identify a bag that can be utilized in the same manner as current bag, but that has increased resiliency following extensive manipulation.
In all, six companies submitted bags for evaluation, and several companies provided more than one type of bag. Suppliers submitted between 3 and 10 bags that had EVA as a major component. The bags had a comparable configuration to the current bag, with several different materials or combination of materials. Supplier E could not provide a bag with the required dimensions, but did have a smaller bag in a comparable material that was submitted for testing.
Methods
This study attempted to handle bags consistent with real world practices. We performed a total of three tests of the bag integrity. The first test provided us with preliminary results which we used to refine the experimental design for tests 2 and 3. In particular, we performed drop tests 2 and 3 at -30C, rather than at room temperature (the condition for test 1), in order to prevent warming of the bags, assuring worst case scenario for film toughness and flexibility for the various types of films.
All bags of comparable dimension were filled with 4.5kg of purified water as a surrogate for the product. The bags from supplier E that were not of comparable dimension were smaller. Because the smaller bag contained less weight, it was less likely to fail. To compensate for this, the smaller bags were intentionally over-filled. No effort was made to equate the degree of overfilling with the performance of the bag.
The bags were frozen and packed in plastic containers (totes) for transportation in a manner consistent with the current process. Normally, these totes are transported by fork truck. The jostling of the bags in totes during transport could be one of the causes of leaks when the bags are thawed, so the bags were transported between buildings as part of the durability testing. Each tote contained 3 bags, and approximately 8 totes were loaded into a protective cage and on to the forklift. Once the forklift arrived at the warehouse, team members unloaded the totes from the forklift. When bags are stored in the warehouse, they are periodically manually inventoried. To ensure these test bags are subject to all typical handling scenarios, we asked personnel to inventory the test batch. Following this exercise, we reloaded the totes onto the forklift and drove them back to the production building.
At the production building, we moved the totes to an unclassified room with floor drains. We placed the bags on the top shelf of a rack used in the production area. This shelf is approximately 5.5 ft off the ground. As a worst case challenge, we pushed the bags off the top shelf one at a time, numbered them, and photographed any damage to the bags.
For tests 2 and 3, we used 5 bags from each manufacturer. The bags were treated the same: filled, transported by forklift, and returned to the building were they were filled. The key change from test 1 was that the bags were moved to the -30C storage site where the drop test was performed for both tests 2 and 3.
Results First Drop Test
We obtained sample freeze bags for evaluation from six supply companies. We tested up to 10 samples of 5-liter bags from the supply companies (range n=3 to n=10). Two manufacturers (C and F) had more than one type of bag available for testing. We were unable to obtain 5-liter bags from supplier E, so instead tested theavailable smaller bags. Each bag was filled and subjected to the worst case challenge of mishandling and drop at room temperature. The order of the bag drop was random across all suppliers. After the drop, bags were completely thawed and results were tabulated. Bags were considered to have survived the test if they did not leak upon thawing. The results of the first trial are captured in table 1:
Only 7 of the 59 bags dropped survived. There were three confounding issues associated with this first test that prevented us from drawing a definitive conclusion about bag reliability.
First, supplier E could not supply 5 liter bags in time for the test. To compensate, we used smaller bags and overfilled them. Two of the five smaller bags survived, which was a 40% survival rate, the highest of any manufacturer. It is possible that the high survival rate was due to the relative strength of the smaller bag. Second, the number of bags available for testing was variable. Third, the order of bag drop was non-systematic. With the exception of supplier E, none of the first 40 bags dropped survived the test. The bags that survived had been at room temperature for a longer period of time, and their contents had begun to thaw. The increased temperature of the bag may have contributed to the survival of the bags. For these reasons, we conducted an additional test of bag durability. We were particularly interested in further testing of the bags from suppliers B and E, as these had the highest pass rates.
Second Drop Test
To control the variable temperature of the first drop test, we conducted a temperature-controlled the experiment. Here, the bags were moved into a -30C freezer after mishandling and transport. Bags were given adequate time to reach an external temperature of -30C. We tested 5 five-liter bags of each sample. 
The order of the bag drop was random across all suppliers. After the drop, bags were completely thawed and results were tabulated. As in the first test, bags were considered to have survived the test if they did not leak upon thawing. The results of the second trial are captured in table 2:
The bag from supplier E was the only sample with more than one bag survival. In order to obtain a better measure of bag durability, we conducted an additional test to increase our sample size.
Third Drop Test
The third drop test was similar to the second test, but with the mishandling and transport portion omitted. Based on our observations during the first two tests, most of the damage occurred during the drop test. We tested 5 five-liter bags of each sample. The order of the bag drop was random across all suppliers. After the drop, bags were completely thawed and results were tabulated. As in the first tests, bags were considered to have survived the test if they did not leak upon thawing. The results of the third trial are captured in table 3:
Again, supplier bag E had was the sole sample with more than one bag survival. We tabulated the results for all 5-liter bags (excluding the smaller bags from the first test). These results are in table 4.
The combined results show the supplier E bag was the most promising bag to survive real-world transport, storage, and handling in the plant. In a -30C freezer, 50% of the frozen bags survived a 5 foot drop. Supplier E provided a bag that consistently outperformed the other bags in all three tests.
Recommendations and Conclusions
Through worst-case challenge tests related to the manufacturing operation, we were able to identify a bag that we predict would perform better in routine operations. Implementing a more robust bag protects against loss due to damaged bags. Switching to a new frozen storage bag has several challenges, since the bags must be compatible with the product, as well as the existing production and warehouse facilities. Qualification of the new bags is necessary to complete the change. Costs of this qualification must compare favorably to the projected cost savings of implementation of the more robust bag.
Acknowledgements
The authors would like to acknowledge the following individuals for their significant technical contributions: Anju Chatterji, Aaron Neilsen, Patricia Kitchen, Jacob Tesfai, Brandon Foster, Joyce Liu, Dale Dwier, and James Flowers
References
1. Wong, Russell, Disposable Assemblies in Biopharmaceutical Production, Bioprocess International, Disposables Supplement, October 2004
2. Sinclar, Andrew and Monge, Miriam, User Viewpoints on Disposables Implementation, BioPharm International, June 1, 2009
David Kilburn is a Conformance Investigator at Bayer Healthcare in Berkeley, California. Since 2000, he has investigated discrepancies in the Active Pharmaceutical Ingredient department. David has a BS in Chemical Engineering from the University of California, Berkeley. Russell Wong is a Senior Process Development Engineer in Manufacturing Sciences at Bayer Healthcare in Berkeley, California. Since 2000, he has been responsible for single use systems technologies development and implementation with a focus on improving reliability and performance of these systems. He worked previously in Bayer Plastics Technology and Monsanto Plastics Technology prior to it’s acquisition by Bayer.
Russell has a PhD in Polymer Engineering from the University of Tennessee, an MS in Bioengineering from the University of Utah, and a BE in Chemical Engineering from the Cooper Union for the Advancement of Science and Art.
Arn Malliett is a Senior Quality Assurance at Bayer Healthcare in Berkeley, CA. Since 1999, he has been involved in discrepancy management within the fermentation and purification process streams. He has also participated the facility startup and implementation of new documentation programs. He has previously worked in Process Development and Manufacturing. Arn has a BS from the University of California Berkeley.