February 8, 2005

Silver Spring, MD



Topic # 21:    : Risk  Risk Assessment Model Assessments  for Potential Risk of Exposure to Variant

t Creutzfeldt-

Jakob Disease (vCJD)AgentvCJD) Agent in Plasma Products


Issue:  FDA seeks the advice of the Committee on the design and input parameters of a risk assessment model for potential vCJD exposures from products made with U.S. plasma.




Background and Rationale


The TSEAC has not been presented with a detailed risk assessment for CJD and vCJD in plasma products since reports of presumed transfusion transmission of vCJD in the United Kingdom, in December 2003 and July, 2004.  Two cases of presumed transfusion transmission of vCJD were reported in the United Kingdom, respectively in December 2003 and July 2004.  The U.K. performed commissioned – through Det NorskeVeritas - an updated risk assessment based upon these new vCJD transmissions.[1]  As a result of theiris assessment, U.K. authorities have informed U.K. recipients of FVIII, through their physicians, that they may be at and some other plasma derivatives that they are at  increased risk for vCJD.[2]   Some recipients of other products were also informed.[3] Correspondingly, FDA seeks to understand the vCJD risk from U.S. licensed products that are made with U.S., as compared with U.K. sourced plasma.  Such estimates may serve as a basis for reexamination of the adequacy of current measures to protect blood and plasma derived products.  Towards that end, FDA staff have developed a risk assessment model for CJD and vCJD for discussion.



Risk assessment is important from the standpoint of determining adequacy of current measures to protect blood and plasma derivatives.  However, Iin the setting of uncertainties, risk assessments cannot predict risk preciselyrovide a precise determination of risk, but rather supply a range of possible risks.  These ranges aremay be refined over time as more scientific and epidemiologic information becomes known.  This risk assessment estimates the likelihood of potential exposure to the vCJD agent, not the likelihood of developing a clinical infection. 


The risk assessment approaches used by the FDA generally follow the four-part framework and guidelines of the National Academies of Science.[4]

An assessment of vCJD exposure risk in U.S. products is expected to provide:


·        A framework for more precise risk assessments in the future

·        Ranking of product classes that may have greater or lesser margins of safety 

·        Estimation of likely, best-case, and worst-case risk of exposure to vCJD via products

·        Estimation of the need for furtheradditional risk reduction measures

·        Estimation of of what levels of TSE clearance in manufacturing is that are likely to be meaningful

·        Risk communication to the public


Variant CJD risk assessments are limited because some of the needed information is not available.  In these cases, the approach is to use a range of estimates based upon current knowledge and expert opinionelicitation.  In the case of vCJD, the prevalence of infection is not known among blood donors in the U.S., but is estimated to be very  low.  The amounts of infectivity in blood of people incubating vCJD is not known, but have been is estimated based upon observations in animal models.  The TSE clearance capacity of manufacturing processesin products is based upon animal model studies,usingstudies, using spiking material that may not entirely reproduce the form of the TSE agent in plasma.  The susceptibility of recipients to infection is also an unknown factor, although people that are heterozygous for methionine/valine or or homozygous for valine at prion protein codon 129 may be at lower risk.  Transfusion transmission of vCJD occurred in a person heterozygous for methionine/valine at that locus suggesting that no genotype may protect absolutely against infection. The discovery that one of the U.K. vCJD transmissions by blood occurred in a heterozygous person casts uncertainty about the relative resistance to vCJD by transfusion in people with this genetic background.


Elements of vCJD Risk Assessment for Plasma Products


            A risk assessment usually consists of a mathematical model and a document detailing the scope of the assessment, the data and information used, the assumptions, methods, results and conclusions. Three main categories of information are needed to perform a vCJD risk assessment:  the amount of infectivity in starting material, the amount of clearance during manufacturing, and the amount of product given to a recipient.  Because of the uncertainties associated with these components a probabilistic model using statistical distributions to represent the information for each variable is used.  Exposure to the vCJD agent is estimated in the exposure assessment of the risk assessment.The likelihood of an infected donor of U.S. plasma is difficult to estimate, but   Important factors for determining the probability and level of vCJD exposure include tThe likelihood of an infected donor of U.S. plasma, although difficult to estimate is unknown, but is likely to be extremely low due to geographic blood donor deferrals to reducefor possible exposure to BSE.  The quantity of iInfectivity is estimated based upon experiments in animals, as approximately 0.5 – 500 intracerebral (ic) ID50/ml of plasma.  The ID50 is defined as the dose that is necessary to initiate infection inaa 50 % of those exposed.  probability of exposure to infectioninfection.  Thus exposure to 1 ID50 would suggest a 50% probability of exposure to of infection, and 0.1 ID50 would suggest a 5% probability ofprobability of exposure to of infection.  In this model, an  ID50  is based on data from animal studies, and the relevance to humans of an ID50 derived from  animal studies is unknown.  The amount of TSE clearance in products is estimated based upon steps in manufacturing that have been shown to result in clearance clearance of spiked or endogenous infectivity in pilot studies with animal models.  Since such studies have not been performed for all plasma derivatives, and there is biological variability associated with TSE animal models and manufacturing pilot studies, the logs of  clearance used for the purpose of this modelrisk assessment are defined as rangesis a range.  can be highly variable or may have not have been performed for all plasma derivatives, the logs Log10 clearance used for the purpose of this risk assessment is a expressed as a range of values.   Each plasma derivative product is unique in its manufacturing processesterms of details of manufacturing, resulting in the likelihood so that identical levels of clearance will notcannot occur among  be assumed for different products.   Published data from animal studies in model TSE’s are used to define the upper and lower vCJD clearance estimates for manufacturing processes. 

RThe results in the risk characterization portion of the risk assessment will beare expressed using a measure of central tendency (e.g. mean, median, etc.) and a confidence interval (usually 95% CI) to express the uncertainty associated with the resulting risk estimate.  A sensitivity analysis in which parameters are varied by a preset percentage of the value (10%, 25%, 50%, etc.) is conducted to determine the parameters that have the greatest influence on the final risk estimate and identifies areas where additional information may improve the risk estimate.  Finally, aFinally, a discussion of uncertainties and data gaps provides useful information for data needs and research priorities.      


Several The specific parameters are critical to generating awere used for the U.S. risk assessment are:


·        Prevalence of vCJD in screened blood donors, with modifiedcations to allow made  for effectiveness of blood donor deferral, and possible exposures to BSE o of short duration to the BSE agent, for which blood and plasma donors are not deferred.

·        Infectivity of plasma from personsa incubating vCJD-incubating person, based upon animal studies.   The assigned numbers may be viewed as represent closer to a maximum possibility, because it may be that some infectedsuch donors would not have infectivity in express blood borne infectivity.

·        Plasma manufacturing pool size (number of donors contributing, volume of donations).

·        TSE clearance by manufacturing processes, based upon published TSE cclearance studies with animal-adapted TSE agents.  Since many products have not fully undergone such studiesy, a range of clearances is selected. 

·        The amount of possible infectivity that might be contained withiin a single dose of product.

·        The amount of product given to a typical recipient per course of treatment or per year. 


Interpretation of even the worst caseworst-case scenario mustshould take into account the many uncertainties inherent in underlying assumptions.




The document “Draft Risk Assessment: Potential Exposure to the vCJD Agent through Human Plasma-derived products manufactured in the United States” evaluates potential exposure to the vCJD agent through human plasma derived Factor VIII product manufactured in the U.S. from plasma pools that may have contained a donation from a vCJD infected individual. An abbreviated summary of the entire document is provided below.  This assessment estimates the probability and level of exposure to the vCJD agent and the accompanying uncertainty of the exposure estimate.  Because scientific data on the level of exposure to vCJD agent and the likelihood of certain human health outcomes, such as infection and illness, are lacking, it is not possible to provide a precise estimate of the vCJD risk to patients potentially exposed to the agent through plasma derived products.


Question and scope addressed by risk assessment: Given the recent transmission of vCJD via transfusion of whole blood and component products in the United Kingdom, what is the risk of potential exposure to the vCJD agent to the U.S. plasma derivative recipients?hemophilia population who have received human plasma derived Factor VIII?


In drafting a risk assessment model, many input variables must be considered. 


Exposure Assessment


The exposure assessment component of a risk assessment evaluates the routes of exposure to a hazard, the probability that exposure occurs and the amount of a hazardous agent to which a person or population may be exposed.  This exposure assessment specifically addresses the probability of exposure,  (and if present,) the quantity of vCJD agent that may potentially be present in plasma derived products manufactured in the United States.  The route of exposure is definedclear,, since plasma derivatives are usually administered to patients either intravenously or intramuscularly.


The assessment assumes that plasma pools consisting of 20,000 or more donations collected from U.S. plasma donors arewere  used,  as the starting material from which a number of plasma- derived products are purified, including Factor VIII.  Because of the relatively large number of donations per plasma pool there is a small probability that even in the United States that some plasmaof the pools may contain a donation from a donor that iswho may  unknowingly be infected with vCJD.  Manufacturing processes may remove and reduce a certain percentage of the vCJD agent in the product.  Ultimately exposure can potentially occur through use of plasma derivedplasma-derived products.  The overview and text that follows describes the components that contribute to potential exposure to the vCJD agent  and the model in more detail.




Overview of Model


Module 1 – Estimation of potential vCJD incidence in the U.S. This portion of the model estimates the annual number of both symptomatic and asymptomatic vCJD cases in the United States. To date, there have been no reported cases of domestically acquired BSE in cattle or human vCJD in the U.S.  Theoretically, cases of vCJD in the U.S. might be expected to arise from two potential sources: (1) dietary exposures to the BSE agent to incurrent U.S. residents with extended that traveled to or residenced in the U.K. or other parts of, Europe or other countries, where BSE occurred, for extended periods of time or, or (2) dietary exposure to the BSE agent via infected cattlecontaminated beef  in the U.S.  Based upon current USDA surveillance data, tThe potential for dietary exposure to BSE in the U.S. donors iswas estimated to be negligible, if any.  .  For any potential vCJD-incubating donors, cases it iswas assumed that the level and infectivity of vCJD agent present in blood and plasma remain at a constant level in the blood of vCJD infected individuals during the entire incubation period.  Therefore, any plasma donation from a vCJD infectedvCJD-infected person is assumed capable of transmitting the disease to a recipient.  


Module 2 - Plasma Donation and vCJD.  This module estimates the percentage of potential vCJD contaminated plasma pools based on the predicted potential annual number of people incubating vCJD cases  among the U.S. population, the rate and frequency of plasma donations, and the total number of plasma pools produced. The model also considersed the level of effectiveness of the donor deferral policies in screening and eliminating potentially infected donors. Potential exposure to the vCJD agent wisas estimated in the model by considering the probability of a vCJDcontaminated  donation being present in a plasma pool and the amount of amount of  infectivity that might be present in the contaminated donation. Given the low prevalence of possible donations from vCJD-infected individuals, the model assumes that if vCJD agent were present at all in a plasma pool, there would likely be only one such donation per pool. 


Module 3 - Manufacturing ProcessFVIII Processing.  The quantity of vCJD potentially present in product FVIII vials made from vCJD-exposedcontaminated plasma pools may be was estimated based on the possible infectivity present in an entire contaminated pplasma pool, and the total yield of product from FVIII from plasma.  As material from a plasma pool iswas fractionated during manufacture it iswas assumed that the amount of vCJD agent present iswas reduced through clearance mechanismsremoval of the agent.  The degree of reduction varies depending on the type of specific fractionation and purification steps used. Plasma derived FVIII products were categorized as intermediate purity FVIII, and immunopurified FVIII (high purity) based on the specific plasma fractionation and purification procedures used during production.


Module 4 - Utilization of ProductsFVIII by Hemophilia A Patients.  Potential eExposure of an individual hemophilia A ppatient to the vCJD agent can bewas estimated in the model based on the total quantity of productFVIII used per year, and the estimated potential quantity of vCJD agent predicted to be present in the FVIII product. The quantity of productFVIII utilized by an individual patient is dependent on the severity of the disease and the treatment regimen of choice. Our model can estimate This risk assessment estimates the annual potential exposure of individual patients to Hemophilia A patient to the vCJD agent. s for those with severe hemophilia A disease treated either on an episodic or a prophylaxis basis.  Those with severe hemophilia represent approximately 50% of hemophilia population in the U.S. and they are presumably at higher risk of potential exposure to the vCJD agent via human Factor VIII.


Model of Exposure Assessment












































Results from the Model

The results from the model indicate that on average approximately 2.5% of Factor VIII vials (95% CI   0.6% - 5.4%) manufactured in the United States and the plasma pools used to manufacture them may potentially contain vCJD agent.  Given the low prevalence of possible donations from vCJD-infected individuals the model assumes that if vCJD agent were present at all in a plasma pool, there would likely be only a single vCJD donation present per pool.  Furthermore, the amount of vCJD present in a plasma pool would likely be reduced during manufacture of the plasma-derived product.  The results from the model given below represent estimates of potential infectivitycontamination predicted to be present per contaminated unit, or vial of human plasma-derived Factor VIII (Table I) if a contribution was made by an infected donor.  Table II provides predictions for annual treatment regimens for Hemophilia A given either on a prophylacticxis or episodic basis to control the occurrence of bleeding.



Note* - Results in Tables I and II are draft results and may vary from final presentations based on peer review input on the risk assessment


Table I – Exposure to vCJD iv ID50 through Factor VIII. Quantity of vCJD iv ID50 predicted by the model to be present per single unit of Factor VIII and per vial of 1,000 u Factor VIII product of either high or intermediate purity.  High purity product is immunopurified product.





Type Factor VIII Product


Quantity Factor FVIII



Mean vCJD

iv ID50











High purity



1 u









Intermediate purity










High purity


One vial of

1,000 u









Intermediate purity








* iv ID50 represents the probability that 50% of those exposed to 1 ID50 intravenously may potentially                            become infected with vCJD.











Table II – Annual Exposure to vCJD iv ID50  through use of Factor VIII for treatment of Hemophilia A patients with severe disease.  Quantity of vCJD iv ID50 predicted by the model in the total quantity of Factor VIII  product used to treat a patient.  High purity product is immunopurified product.





Type Factor VIII Product



Regimen for Severe Disease




Utilized     per Year


Mean vCJD

iv ID50












High purity




236,800 u(1)











95,200 u(2)











Intermediate purity





236,800 u(1)











95,200 u(2)









* iv ID50 represents the probability that 50% of those exposed to 1 ID50 intravenously may                                      potentially become infected with vCJD.

(1) Prophylactic use 5th and 95th percentiles are 158,800 u and 314,600 u, respectively.

(2) Episodic use 5th and 95th percentiles are 63,800 u and 126,400 u, respectively.

(3) Because the mean percentage of vCJD-contaminated vials is approximately 2.5% the 5th (not shown) and 95th percentiles would be expected to be zero, therefore results for the 99th percentile are presented.



Sensitivity Analysis


Sensitivity (or importance) analysis is a process of varying the value of variables in the model to identify those with the greatest influence on the estimated risk outcome(s).  Such variables include the estimated incidence of vCJD in the U.S., which is dependent on factors such as the incidence of vCJD in the U.K., and the effectiveness of donor deferrals for vCJD risk.  Other important variables A simple sensitivity analysis of the Factor VIII model suggests that the estimated incidence of vCJD in the U.S. (PvCJD ) had the greatest influence on the final risk estimate, include the total number of  donations in a pool, and followed by the second factor - the number of total donations per pool (Dpool).  Another influential factor in the model was the log reduction of vCJD agent (RLog) during processing and manufacture of products of Factor VIIIXI product.  For RLog it was assumed that the minimum level of reduction was 0, and the maximum level of reduction in agent that could occur was 7 log10.


Uncertainty and Data Gaps


Uncertainty arises from the absence of information or availability of limited information.  In our probabilistic model statistical distributions can beare used to represent the uncertainty of the information used in the model.  We express the Uuncertainty of the final risk estimates generated from the model can be expressed asusing a a mathematical mean (average) of exposure in ID50 units and the 5% and 95% confidence intervals for each estimate.  The uncertainty for the risk estimates generated from the FDA plasma derivatives risk assessment model is significant and decision makers should use the results with caution.

.  In the future, additional research and information may be substituted for assumptions or used to improve estimates for the individual parameters and ultimately improve the precision of the final risk estimates generated by the model. 


Many of the variables and parameters used in the model have significant uncertainty associated with them – in some cases expert opinion and assumptions were used to estimate the values and ranges of certain parameters used in the model which will add to the uncertainty surrounding the final risk estimates. In the future, additional research and information may be substituted for assumptions or used to improve estimates for the individual parameters and ultimately improve the precision of the final risk estimates generated by the model. 


Even considering the associated uncertainty of estimated risks, risk assessment provides a best estimate of risk based on the current and known information.  It is still a useful tool that informs the science-based decision making process.  It can identify data gaps and research priorities where additional research and information would have the greatest impact on enhancing the final risk estimates.  Results from the sensitivity analysis in the previous section indicated that estimation of the incidence of vCJD in the United Kingdom, and in turn, the parameter for vCJD donations per plasma pool (PvCJD ) had the greatest influence on the final risk estimate.  Improved data and surveillance studies on the estimated vCJD incidence in the U.K. would enhance the precision of the estimates for this variable and the resulting risk estimates from the model.  Likewise, better data on the level of log10 reduction of vCJD agent (RLog) during processing and manufacture of Factor VIII product would decrease the uncertainty.  Again, more precise information would improve estimates from the model. 




No data are available on the level of infectious units or ID50 units present in the bloodstream of vCJD infected individuals at the time of blood donation.  The model extrapolates an estimate of the level of vCJD agent that might be present in human blood based on data from several animal models.  However, the level of agent present in an infected individual at the time of blood donation could differ from our assumption and this adds to the uncertainty of the risk assessment outcomes. 


The model estimates exposure to the vCJD agent in the form of intravenous ID50 units.  Data are not available to estimate the probability of various clinical outcomes, such as infection or illness that might be predicted to arise from exposure to a particular level of agent.  Therefore, the model will not estimate a probability of infection or illness.we did not estimate a probability of infection or illness in our model.  However, aA meaningful dose-response model will need to be generated for vCJD exposure in humans to estimate the probability of adverse clinical outcomes for humans.  Until then, estimates of the probability of vCJD infection or illness arising from exposure to the agent remainare extremely uncertain. and should be viewed with caution.





Potential exposure to the vCJD agent present in Factor VIII manufactured from human plasma in thein plasma derivatives manufactured from U.S. plasma United States was can be modeled.  estimated in this risk assessment. The selection of parameters for the model is likely to strongly influence results.    


It is likely that only a small percentage of Factor VIII vials produced annually in the United States may be manufactured from plasma pools that may have contained a plasma donation(s) from an individual that may have been infected with vCJD.


Because of manufacturing steps and variation in the quantity of product utilized, the potential vCJD exposure risk for various plasma derivatives maycan differ significantly.  In general, products manufactured by processes that afford limited clearance of the vCJD agent and that are used in larger quantities pose a larger potential for exposure to the vCJD agent. 


Although exposure may occur it is not possible to provide a precise estimate of the vCJD risk to patients that may have used Factor VIII manufactured in the U.S. 








Potential Risk of Exposure to vCJD ID50 (U.S.)


[Description of analysis – summarized, including sensitivity]





Table:  Potential Exposure to vCJD in ID50/person/year



Product  Quantity Used    Mean vCJD ID50    5th %ile  95th %ile



(low purity)





high purity







Questions to the Committee


            FDA requests the committee’s consideration and comment on the U.S. risk model. 


Please comment with regard to:


a.       The model per se; and

a.b.  Any additional information that is needed to improve risk estimates for the various plasma derivatives



[1] Det Norske Veritas Risk Assessment 2003, at

[2] Health Protection Agency website, at

[3] See attachment 1, vCJD and plasma products: who may be affected?  Ibid, under “Further Information.”

[4] Risk Assessment in the Federal Government: Managing the Process. 1983.  Washington DC, National Academy Press