Before a pharmaceutical product is introduced into humans, either in a clinical trial or as a marketed product, virus safety¬†must be evaluated carefully. Virus¬†safety normally is ensured using a
three step complementary approach:¬†selecting and testing cell lines and/or¬†raw materials for the absence of¬†viruses, testing the product at¬†appropriate steps of production, and¬†assessing the capacity of a production¬†process to clear infectious viruses (1).¬†The latter (also referred to as viral¬†clearance) is the subject herein. Spiking¬†studies are conducted to evaluate the¬†capacity of a purification process to¬†clear viruses. Steps that are expected¬†to clear viruses are studied with¬†down-scaled laboratory experiments,¬†in which different types of viruses are¬†introduced deliberately and virus¬†content before and after the step is¬†evaluated. Clearance of a virus is¬†expressed as a reduction factor (RF)¬†calculated as the total virus before a¬†virus clearance step divided by the¬†total virus after the step. It is often¬†expressed logarithmically as a log¬†reduction value (LRV) and calculated¬†as LRV = log10 (RF) (2).
When virus clearance of different¬†purification steps takes place by¬†different mechanisms (e.g., size¬†exclusion and inactivation by¬†detergent), virus clearance of such¬†steps is considered to be additive. By¬†characterizing several steps, total virus¬†clearance can be calculated by adding¬†individual LRV values. Total virus¬†clearance capacity of a purification¬†process should be substantially higher¬†than the potential virus load in the
starting material for purification.
However, virus clearance can be¬†compromised if a process step with¬†potential high levels of virus cross-contaminates a later purification step¬†(carryover). This might occur when¬†operators use the same equipment¬†(e.g., a pH probe or transfer pump)¬†before and after a virus clearance step¬†without effective cleaning that¬†equipment or when aerosols are¬†created in different process steps in¬†open processes in the same room.
To assess the potential impact of¬†carryover from an early purification¬†step to a later one, we have developed¬†a simple equation for calculating and¬†identifying a critical potential carryover¬†(CPCo), which is defined as the¬†volume of carryover that will¬†significantly affect the overall virus¬†clearance of a purification process.¬†Based on this calculation, mitigations¬†can be introduced to prevent such¬†carryovers.
Biomanufacturers should not accept¬†a carryover risk that can be prevented¬†by complying with good manufacturing¬†practices (GMPs). Biopharmaceutical¬†products for human use always should be processed according to GMPs,¬†which state that appropriate precautions¬†should be taken to prevent viral¬†contamination from before to after¬†virus removal and inactivation steps (3).
Calculation of Critical Potential Carryover
To evaluate CPCo of a virus-containing material during a
manufacturing process, the CPCo
volume can be calculated. CPCo is
defined as the maximum volume of a¬†potential carryover that would not¬†compromise significantly the overall¬†virus safety of a final product.
The virus RF of a purification step¬†is calculated as the total amount of¬†virus before that step divided by the¬†total amount of virus after that step¬†(2). If carryover happens during
production, virus content in the¬†output from a step increases and¬†clearance decreases because of the¬†amount of virus present in the cross-contaminating material. True virus
reduction factor (RFtrue) can be¬†calculated according to Equation 1.
The amount of virus content before¬†the step = Tbefore¬†√ó Vbefore, where Tbefore
is the titer of virus in the material¬†before the step (virus particles per¬†milliliter), and Vbefore
¬†is the volume of¬†material before the step.
The amount of virus after a process¬†step is calculated according to¬†Equation 2, where Tafter¬†is the titer of¬†virus in a material after the process¬†step (virus particles per milliliter),
Vafter¬†is the volume of material after¬†the step, VCo¬†is the volume of¬†carryover, and RFlab¬†is the reduction¬†factor found in the laboratory for that¬†particular process step.
Equation 1 shows that if there is¬†no cross contamination or carryover¬†(VCo¬†= 0), then RFtrue¬†= RFlab. It also¬†shows that the potential cross-contamination will not constitute a
problem to the final virus safety if¬†VCo/Vbefore¬†is much lower than 1/RFlab.¬†So to calculate CPCo, a maximum¬†acceptable value of VCo/Vbefore¬†in¬†comparison with 1/RFlab¬†must be
Because of the nature of virus¬†clearance, an amount of virus may¬†remain after a virus clearance step.¬†This amount is determined by the¬†amount before (Tbefore¬†¬†√ó Vbefore) and the¬†reduction factor. If the amount of¬†virus introduced by carryover¬†contamination is of the same¬†magnitude as that remaining after a¬†given step, then RFtrue¬†= RFlab/2,¬†which happens if VCo/Vbefore¬†= 1/RFlab.
Reduction of RF by a factor of two¬†corresponds to a reduction in LRV by¬†0.3. Because a typical standard¬†deviation for determination of LRV for¬†a single virus clearance step is in the¬†order of 0.5 log10, a deviation of 0.3¬†log10¬†will not change dramatically the¬†overall evaluation of virus clearance for¬†a purification process, which typically¬†consists of several virus clearance steps.¬†The volume of a CPCo may be defined¬†as: CPCo = VCo¬†= Vbefore/RFlab.
Based on that, CPCo can be¬†calculated between different process¬†steps, and it can be used for¬†determining carryover volume for a¬†series of successive steps. When those
volumes have been defined for a given¬†purification process, this information¬†can be used to focus on where in a¬†process different segregation measures¬†must be applied. Therefore, such¬†mitigations can be made based on¬†process specific knowledge and
segregation measures in a facility.
Possible Sources of Carryover
To further evaluate the necessity of¬†possible mitigations, different types of¬†potential carryover should be¬†identified.
Aerosols: Aerosols can have a¬†diameter from 0.01 to 10 ¬Ķm. Here it¬†is assumed that aerosols that can easily¬†flow in a production room have a¬†diameter of about 0.01‚Äď0.1 ¬Ķm (4). An¬†aerosol with a diameter of only¬†0.01 ¬Ķm (10 nm) will not be big¬†enough to carry a single virus particle,¬†but an aerosol with a 0.1-¬Ķm diameter¬†will be able carry virus particles. A¬†sphere with a 0.1-¬Ķm diameter has a¬†volume of about 5 √ó 10-16¬†mL. One¬†liter of air may contain up to 109¬†particles (4), so even if individual¬†droplets have a very small volume,¬†transfer of material, including viruses,¬†is possible.
Aerosols can be created in a¬†number of scenarios. For example, in¬†an open container with foam, aerosols¬†can be created when bubbles in the¬†foam burst, and aerosols can be¬†formed when the pressure in a¬†pressurized vessel is released. In a¬†worst-case situation for a 1-m3¬†tank,¬†1,000 L of air with aerosols could be¬†created even though the pressurized¬†vessel is equipped with a filter through¬†which the air is released (vent filter).
Biomanufacturers typically expect that¬†only a very small fraction of the¬†aerosols released to the air in a room¬†will end up in an intermediate from a¬†later process step. If 0.1% of aerosols¬†should end up in an intermediate at a¬†later step, the volume of those aerosols¬†can be calculated as Vaer¬†¬†= 5 √ó 10‚Äď16¬†mL √ó 109/L √ó 0.001 √ó 1,000 L = 5 √ó
10‚Äď7¬†mL. Consequently, the volume¬†carried by aerosols will be expected to¬†be in the nanoliter range.
Drops: If equipment that is used¬†before a virus clearance step is reused¬†later without proper cleaning before¬†reuse, it could transfer small volumes¬†of virus-contaminated liquid. This¬†could be the case for pH or¬†conductivity electrodes. Potential¬†carryover of small volumes also could¬†happen unintentionally by the actions¬†of personnel who work both before¬†and after a virus clearance step (e.g.,¬†with a contaminated glove). Such
carryover might be a few droplets. The¬†volume carried by those drops will be¬†expected to be in the microliter range.
Severe: Higher volumes of¬†carryover could happen if large¬†equipment such as vessels or closed¬†production systems (e.g.,¬†chromatography systems) are used¬†without proper cleaning. It also could¬†happen if open vessels are placed near¬†each other, with risks for larger¬†volumes of carryover present. Larger¬†spills from before to after a virus¬†clearance step often can be identified,¬†and the intermediate product can be¬†discarded. A severe carryover can be¬†in the range of milliliters to liters.
Example of Calculating CPCo
As an example, Figure 1 shows a¬†manufacturing process for a¬†monoclonal antibody (MAb) with¬†volumes at different steps and virus¬†clearance values (LRV) for some¬†manufacturing steps. Using the¬†volumes and LRV values in Figure 1,¬†the CPCo for different virus clearance¬†steps can be calculated. For example,¬†CPCo for a low pH step is calculated¬†as VCo¬†= Vbefore/RFlab¬†= 100 L/106¬†=¬†0.1 mL. For all three virus clearance¬†steps, VCo¬†can be calculated as¬†VCo¬†= Vbefore/RFlab¬†= 100 L/1016¬†=
1 √ó¬†10‚Äď11¬†mL .
Calculations in Figure 1 show that¬†carryover of a volume corresponding to that one drop will not compromise¬†the clearance obtained by one step¬†with a LRV in the order of ‚Č§5 log. But¬†a volume of carryover larger than¬†1 mL will compromise each virus¬†clearance step in the manufacturing¬†process in Figure 1. Two virus¬†clearance steps that together
contribute an LRV of >10 log will be¬†compromised by a carryover of even¬†tiny volumes in the order of ‚Č§0.01 ¬ĶL,¬†and three or more steps might be¬†compromised by small amounts of¬†aerosols.
Determining the Right Approach
Effective virus clearance and¬†appropriate segregation of virus-reducing steps are critical elements in¬†the virus safety of biopharmaceuticals.¬†Segregation overkill will lead to
excessive production costs and can¬†lead to prolonged production times,¬†whereas inadequate segregation will¬†compromise virus clearance and thus¬†the virus safety of a final product.
Therefore, segregation¬†considerations should be considered as¬†part of virus safety, and appropriate¬†mitigations must be in place to protect¬†the virus clearance determined in
laboratory-scale experiments. By¬†defining critical carryover volumes,¬†biomanufacturers can determine¬†which mitigations are appropriate¬†under a range of manufacturing
process conditions. In that way, cost-effective manufacturing can be¬†achieved while maintaining effective¬†segregation strategies.
We have derived a simple equation¬†here to calculate the consequence of¬†cross-contamination or to determine¬†critical potential carryover volumes.¬†As a way of performing a quantitative¬†risk assessment, we have defined a¬†critical potential carryover volume as¬†the carryover volume that will cause a¬†0.3 log reduction in virus clearance for
an individual process step or for a¬†series of steps. That information can¬†be used to evaluate which types of¬†mitigations are appropriate at different¬†stages of purification procedures. That¬†method does not endorse poor GMP¬†practice and deliberately poor
segregation. It can, however, be useful¬†for illustrating the challenges of¬†maintaining LRV claims and for¬†getting quantitative estimates of the¬†impact of cross-contamination in
1 ICH Q5A(R1): Viral Safety Evaluation¬†of Biotechnology Products Derived from Cell Lines¬†of Human or Animal Origin. International¬†Council on Harmonization of Technical
Requirements for the Registration of¬†Pharmaceuticals for Human Use: Geneva,
2 CPMP/BWP/268/95 Note for Guidance¬†on Virus Validation Studies: The Design,
Contribution, and Interpretation of Studies¬†Validating the Inactivation and Removal of
Viruses. European Medicines Agency: London,¬†United Kingdom, 1996.
3 Guidance for Industry, Q7A Good¬†Manufacturing Practice Guidance for Active
Pharmaceutical Ingredients. Food and Drug¬†Administration: Rockville, MD, 2001.
4 Tu KW, Knutson EO. Indoor Radon¬†Progeny Particle-Size Distribution¬†Measurements Made with Two Different¬†Methods. Radiation Protection Dosimetry 24¬†(1/4) 1988: 251‚Äď255.
Corresponding author Per K√¶rsgaard is a senior development scientist in the Virology department at Novo Nordisk A/S, Hagedornsvej 1, HAD 3.163 DK-2820 Gentofte, Denmark; email@example.com. S√łren Kamstrup and Erik Halkj√¶r are principal scientists at Novo Nordisk A/S.