Our recent paper “Overestimation and underestimation of bioburden resistance in radiation sterilization with low average bioburden”, focuses on how verification dose methods perform when average bioburden levels are very low. While these methods remain widely accepted and effective in many scenarios, the analysis shows that their underlying assumptions do not always hold under low bioburden conditions.
The Shift Toward Low Bioburden
This issue is a direct result of how products are manufactured today.
Improved environmental controls, cleaning processes, and production standards have driven bioburden levels down across many product categories. In these environments, microorganisms are no longer distributed in a predictable way.
Instead, bioburden becomes intermittent across samples, highly variable between units, and is often absent in many test samples. This type of distribution is fundamentally different from what verification dose methods were designed to evaluate.
Where the Model Starts to Break
Verification dose methods such as VDmaxSD and Method 1 rely on statistical assumptions about how bioburden is distributed on average across a sample population.
At low bioburden where some samples do not have bioburden those assumptions begin to fail. Rather than reflecting an average, very low bioburden occurs as the presence or absence of bioburden on samples in the verification dose experiment. In statistical terms, very low bioburden appears as an attribute rather than a variable.
Underestimation of Resistance
When only a small number of organisms are present, the outcome of the verification dose experiment is influenced more by the number of samples with bioburden than information about the resistance of the microorganisms. Since the negatives in the test of sterility following a verification dose experiment can be a combination of samples that had bioburden inactivated by the verification dose and samples that never had any bioburden prior to verification dose exposure, it becomes possible to pass the verification dose experiment even in the presence of infrequent but highly resistant microbes.
The potential issue here is that for very low bioburden and lower sterilization doses in VDmax15 or VDmax17.5, the probability of a unit being non-sterile after processing can be higher than expected.
Overestimation of Resistance
Current methods for selecting verification doses in ISO 11137-2:2013 are based on an equation that looks at the number of colony forming units (CFU) prior to verification dose exposure and calculates a theoretical verification dose to produce approximately 1-in-10 positives in the test of sterility following the verification dose experiment. When average bioburden is already at 0.1 or lower in VDmax the verification dose tables in ISO 11137-2:2013 have a verification dose of 0.0 kGy. For bioburden lower than 0.1 the equation used to construct the verification dose table calculates a negative verification dose.
There are a few problems created by using the verification dose equation with bioburden of 0.1 or less:
- A verification dose of 0 kGy has the same probability of passing or failing regardless of the sterilization. The level of risk associated with a 15 kGy or 35 kGy terminal sterilization dose is not equivalent and the current standard does not recognize this.
- The lowest number of bioburden that can be on a single unit of product prior to irradiation is 1 CFU, so this should be the lowest number that is considered for sterilization dose establishment or substantiation.
The problem created by using an unrealistically low verification dose is that the probability of positives in the dose audit increases as the average bioburden decreases below 1. If the statistical model recognized the reality that the lowest nonzero bioburden is 1 CFU, the probability of failing the dose audit would decrease asymptotically towards zero as average bioburden decreases.
Why This Matters Now
Further back in history many products were sterilized at 25 kGy without a lot of examination into what dose was really necessary to inactivate the bioburden on a particular product. This has changed over time as new products and materials emerged which required lower sterilization doses to maintain their functionality, such as many biologics, tissues, resorbable materials, and other radiation-sensitive items.
Some forward thinking manufacturers have also reevaluated the sterilization dose as a means of increasing sterilization capacity – it’s possible to sterilize more products at 20 kGy than 25 kGy in the same amount of time. With recent challenges related to the availability of cobalt-60 for gamma sterilization and the steady growth rate of medical device manufacturing, more companies are looking into more modern and scalable solutions such as lower sterilization doses or switching from gamma to E-beam or X-ray for securing capacity.
Low average bioburden historically has punished manufacturers in other ways. In the 2006 version of ISO 11137-1, manufacturers with average bioburden of less than 1.5 CFU had to perform bioburden testing on a monthly basis. This had a counterproductive effect of punishing manufacturers for having clean processes with low bioburden. This requirement for monthly testing was removed after about 20 years with the updated 2025 version of ISO 11137-1.
The current methods for dose setting in ISO 11137-2:2013 still punish manufacturers for having low bioburden during the dose audit, due to verification doses being set based on average bioburden of less than 1 CFU, as this increases the likelihood of unnecessarily failing the verification dose experiment in a way that is not reflective of the radiation resistance of real products and microbes.
ISO 11137-2 is currently open for revision at ISO so this is the time to get involved and work with your national standards body (AAMI in the U.S.) to make changes in the standard.
Suggested Resolutions
Ultimately a solution to this problem will be dependent on what manufacturers and regulators can agree on. We predict that a discussion on this subject would include the following
Set 1 CFU as the lowest verification dose to be used with VDmaxSD or Method 1 radiation sterilization validations
Consider increasing the sample size for low bioburden VDmax methods at 15 and 17.5 kGy to 20 or 30 samples or more to reduce the probability of passing the verification dose experiment in the presence of infrequent but highly resistant bioburden
Perform microbial identification to genus level on positives in the test of sterility following a verification dose experiment and use this information to investigate and improve control on the source of highly resistant microorganisms if found in device components or the manufacturing in process materials
Improved microbial characterization and identification of positives in the verification dose experiment align incentives between manufacturers having good control over the manufacturing process and the extent of testing.
Additional Articles We Think You Might Like
Have a question? Speak with a sterilization expert today, at your own convenience.
