Thoughts on the concept of microbiological risk assessment for low bioburden pharmaceuticals

For this newsletter things take a slightly abstract and philosophical turn as we consider microbiological risk assessment in the pharmaceutical context. Here, patient safety is always our foremost concern, and the general focus is with non-sterile pharmaceuticals. Some of the concepts have been drawn from a conversation I had with a food microbiologist, drawing a risk symmetry between two industries.

Microbiological risk assessment is a process that should be objective and based on the best available scientific understanding and the output must be presented in a transparent manner. It is, however, packed with uncertainty. The risk assessment for microbiological hazards can only provide risk managers with a “best estimate” of the risk. The basis of this best estimate, whether the average risk (mean), or the most likely risk (mode), or some other metric, needs to be clearly communicated and include a description of why that metric is the best measure of risk, so that a meaningful decision can be reached.

Where we can express this, then an expression of uncertainty and a quantification of variability should be stated. These are key terms for our understanding and for estimating the level of risk (1). Uncertainty should be stated as at least minimum -most likely -maximum e.g. in terms of microbial growth. This is slightly different to variability, for which mean and standard deviation are often sufficient.

Often, when we discuss risk assessment, we are often talking in relatively abstract terms. While this is inevitable, we should ensure that our risk assessments are as grounded as possible in real-world scenarios and situations (2).

Process of risk assessment

The process of risk assessment, as I am sure many are familiar with, concerns hazard identification and then an assessment of each individual hazard. Upon identification, each hazard needs to be characterized in some way (this could be by classification, such as chemical hazards or biological hazards). Microbial hazards can include organisms or toxins produced by microorganisms. Whether an organism is likely or unlikely in a given context depends on the origin and potential transfer routes. This part of the risk assessment process is sometimes referred to as ‘risk characterization’.

Following this, the degree of severity needs to be weighed up (should the hazard exits or occur) and then the likelihood (or exposure) of a process / environment to the hazard considered. The output can be assessed qualitatively or quantitatively, depending upon the risk framework deployed.

Risk assessment can be based on scientific tools (or perhaps tools with a scientific basis) although the outcomes is not necessarily, in itself, scientific. The purpose of the risk assessment process is to provide rational (or logical) and objective assessment of what is known about a microbiological and its origin and effects at a particular point in time. With effects, this is dependent upon a range of factors including temperature, humidity, water activity, available nutrients etc. and, should sufficient growth occur, the patient population (and then in relation to the level of the drug administered to or consumed by the patient) (3).

No risk assessment process is exact, and there are often differences of opinion (especially as the assessment drifts into a qualitative dimension). If possible, generated data should be such that uncertainty in the risk estimate can be determined. With microbiology there is invariably uncertainty, and this should be acknowledged wherever there are ranges or where chance plays a part. Consequently, the risk estimate will include a level of uncertainty. This should be expressed in the risk report including acknowledgement of where and why such uncertainty arose.

Uncertainty arises in many ways, including uncertainty about what is happening in the exposure pathways resulting in human illness; uncertainty about processes that lead from ingestion, inhalation, or intravenous administration through to infection and illness; uncertainty in the factors that dictate the severity of the illness in different people; uncertainty about population and repopulation (such as operator shedding); uncertainty about survival in a cleanroom; and uncertainty about the parameter values that would describe those pathways and processes.

Product types

For sterile pharmaceuticals, the presence of any organism capable of growth presents a problem. For non-sterile pharmaceuticals, we need to consider the dose–response relationship. This can extends to understanding ill-health effects like diarrhoeal illnesses, hospitalizations, skin reactions, and deaths. We may need to divide the patients or consumers into different subpopulations, such as neonates or immunocompromised individuals.

Non-sterile pharmaceuticals / low bioburden processing

Understanding the effects of a hazard are arguably greater when we deal with non-sterile pharmaceuticals (in contrast, unpicking the origins and transfer routes are arguably greater with sterile pharmaceuticals). With non-sterile formulations, we need to understand the dynamics of microbial growth, survival, and death in within our medicine and the complexity of the interaction (including sequelae) between human and drug following administration (as well as the potential for further spread) (4).

Sequelae is used to define any complication or condition that results from a pre-existing illness, injury, or other trauma to the body.

With non-sterile pharmaceuticals there are a number of questions we need to consider based on the recovery of microorganisms and when we have assessed the species and population:

• What is/are the microbial hazard(s) of concern associated with specific pharmaceutical in question?

• What are the physical and chemical properties of the pharmaceutical product matrix may affect the hazard’s survival and persistence in the drug?

• Is the hazard of concern to consumers / patients and what is the likelihood of the hazard causing an adverse health effect?

• What is the specific population at risk?

• Including sensitive populations, acuteness of the illness (acute versus chronic disease) and other complications such as long-term sequelae.

• Sensitivity relates to the hosts’ immune system, the virulence/potency of the microbes and level of exposure to the organisms.

• Sometimes, young children and the elderly are more sensitive to microbiological infection compared to young healthy adults based on the functionality of their immune system.

• What is the scientific evidence, including illness, that this microbial hazard poses a potential risk in the drug product? This may require an understanding of the dose-response relationship.

• What adverse health effects could be associated with the exposure to the hazard and through what mechanisms?

• What host factors and life stages could affect the type and severity of adverse health outcomes among the population at risk?

• How do common exposure pathways link the adverse health effects with the hazard?

• To what extent do environmental conditions affect the hazard’s transfer and fate along the exposure pathway?

Norms and variances

A common starting point with a risk assessment is assessing the existing level of risk. This is sometimes termed the baseline risk, that is the level of microbiological safety risk posed to the pharmaceutical product without any changes being to the current system of control. Such a baseline enables any improvements or intervention strategies to be evaluated.

By understanding the current level of risk as a baseline, this enables us to appreciate the magnitude of a risk after a change has been made relative to this baseline (such as the implementation of a risk mitigation step). In practice it is not always necessary to determine a baseline risk to evaluate proposed mitigations; however, it is invariably useful to do so.

Qualitative vs quantitative

Most microbiological risk assessments end up as being qualitative. Quantitative risk assessments tend to be better suited for situations where mathematical models are available to describe phenomena, such as dose–response models; and where data are available to estimate the model parameters. When theory or data are lacking, then the qualitative risk assessment approach is more appropriate.

In general, hazard identification is largely a qualitative examination of the product hazard and associated potential adverse health outcomes due to specific drug exposure. This is supported by a critical review of knowledge about the hazards and/or pharmaceutical in question. Generally, the term hazard encompasses any microbiological agent able to cause harm, including bacteria, viruses, parasites, fungi, algae, including their toxins and metabolites, as well as prions/

Best case vs worst case

It is often useful to evaluate the best- or worst-case scenarios to get a sense of the most optimistic and pessimistic risk estimates. These scenarios may be used as a filtering technique or as part of a risk profile. This includes deploying the worst-case scenario to filter out whether a risk, or an exposure pathway, is worth a concern. Often no further analysis is necessary if the most pessimistic estimate shows the risk level to be below some threshold of interest, although this needs to be balanced against the drive for continual improvement (5).

Where we are using microbial data, best- and worst-case scenarios are a form of quantified uncertainty using values that include extremes. With an extreme values from an uncertainty distribution, parameters such as the 1st or 99th percentile are useful, provided our data sets are sufficiently large. Where there is uncertainty given organisms, certain genera can be selected where these pose a risk to the product, process or patient.

When weighing up worst and best cases we need to be mindful of compounded conservatism, which occurs when using conservative, upper-bound estimates for multiple input variables in a risk assessment. This can lead to an overestimation of the overall risk. The overestimation is pronounced when dealing with complex systems, such as where several variables are involved in the risk calculation. This can result in a risk assessment that is more conservative than intended. 

Linking environment to product contamination

There is rarely a direct, concrete connection between environmental control and its related monitoring and what is recovered from a product as it is being manufactured, although sometimes there are connections (such as the recovery of certain Gram-negative organisms and recovery of the organisms or toxic by-products within the product). We can present this relationship, in terms of whether or not risk exists, by understanding patterns of prevalence (such as from environmental monitoring, water testing etc.) and bioburden recoveries (concentration or detection). This gives an overall assessment of the risk of exposure. With exposure, this is not only based on the presence of a hazard but with understanding the mechanism or vector by which the hazard can be found on product contact items or enter the product (via physical transfer, ingredient addition, or air stream).

Both prevalence and number can change as the pharmaceutical product is further processed, and as time elapses. Prevalence and concentration are related, and this also needs to be considered. The detection of a low number of organisms may or may not present a higher risk, depending upon the size of a unit (such as the quantity added), albeit that microbial contamination is not, often, normally distributed.

Here there is a relationship:

Such models can be useful, when they based on known properties and can be used to infer future scenarios.

To carry out such an exposure assessment correctly, we need to have knowledge of:

• Each manufacturing stage.

• And have sufficient data, knowledge, time and expertise to allow each stage to be considered.

Once we understand risk from ‘exposure’ we can design our processes to protect product from the exposure.

Models to consider include:

• Event Tree: This maps out scenarios from a contamination event to a defined endpoint of the assessment, e.g. patient use. The aim is to identify the most likely pathways that lead to contamination. The model can also identify variables in need of further data or modelling. Fault Tree: This starts with the occurrence of a hazard and describes the events that must have occurred for the hazard to be present. This can be useful to determine the likelihood of an event by determining the underlying conditions or events that would allow the given event to occur.

• Dynamic Flow Tree: This model emphasizes the dynamic nature of bacterial growth and incorporates predictive microbiology using statistical analysis of data.

• Modular Process Risk Model (MPRM): This approach divides each stage into microbial and product handling processes. With microbial, this includes where can growth occur and what steps are in place for removal or inactivation. The product handling processes include mixing of units, partitioning of units, removal of parts of units and cross-contamination of organisms among units. A unit is a physically separated quantity of product in the process.

Inactivation or removal are based on the severity of the treatment to the microorganism and its duration.

• HACCP: Hazard analysis and critical control points approaches can be useful for supporting other models, identifying where hazards are likely to occur from product flow models. This approach can also assist in assigning suitable monitoring locations.

Each of these models should be approached with an understanding of dependencies and variables.

MPRM

MPRM presents an interesting model. It can be explained by considering its core elements:

Modularity: The core idea is to divide the pharmaceutical production process into smaller, manageable modules. 

Basic Processes: These modules typically represent fundamental processes like growth, inactivation, mixing, partitioning, removal, and cross-contamination. 

Quantitative Microbial Risk Assessment (QMRA): MPRM is a tool used in QMRA, which aims to quantify the risk of contamination by assessing the probability and severity of microbial exposure. 

Exposure Assessment: MPRM focuses on the exposure assessment stage of risk assessment, which involves estimating the number of microorganisms present in the environment and in a sample. 

Scenario Analysis: The modular structure allows for easy comparison of different scenarios and understanding how changes in one module might affect the overall risk.

Benefits: MPRM provides a structured approach for risk assessment, facilitates the identification of critical control points, and supports risk management decisions. 

Compared with HACCP, MPRM utilizes probabilistic modelling and simulation, including Monte Carlo simulation, to understand the spread and impact of hazards. In contrast, HACCP focuses on identifying and controlling Critical Control Points (CCPs).

References

1.     Anderson, E.T. and Hattis, D., 1999. Uncertainty and variability. Risk Analysis 19: 47-49

2.     Cassin, M.H., Paoli, G.M., and Lammerding, A.M., 1998b. Simulation modeling for microbial risk assessment. Journal of Food Protection 1560-1566

3.     Cerf, O., Davey, K.R., and Sadoudi, A.K., 1996. Thermal inactivation of bacteria- A new predictive model for the combined effect of three environmental factors: Temperature, pH and water activity. Food Research International 29: 219-226

4.     Baranyi, J. and Roberts, T.A., 1994. A dynamic approach to predicting bacterial growth in food. International Journal of Food Microbiology 23: 277-294

5.     McMeekin, T.A., Olley, J.N., Ross, T., and Ratkowsky, D.A., 1993. Predictive Microbiology: Theory and application. Wiley, New York