Mr. Chairman and members of the subcommittee, good morning, and thank you
for inviting me to speak to you about the potential risk of transmitting
Creutzfeldt-Jakob disease (CJD) via blood or blood Products. My name is Dr. Paul
Brown; I am a Board Certified Internist who has spent most of his Professional
career at the National Institute of Neurological Disorders and Stroke
investigating various aspects of a group of diseases known as the transmissible
spongiform encephalopathies, of which CJD is the best known example.
CJD is a fatal degenerative disease of the brain that mainly affects adults
between the ages of 50 and 75 years. It usually begins with a loss of memory
that gradually progresses to frank dementia, and is accompanied by physical
deficits such as incoordination, slurred speech, visual loss, muscle twitching
(myoclonus), rigidity, weakness, mutism, and coma. The entire process is not
unlike Alzheimer's disease run in "fast forward," and plays out to an invariably
fatal ending in less than a year's time from the onset of symptoms.
Left to its own devices, CJD afflicts only one in a million people each
year, which translates to about 250 cases in the United States' a figure that
exceeds our present day experience with polio or rabies but falls short of the
concerns presented by AIDS, viral hepatitis, herpes infections, or even measles.
Why should this comparatively rare disease be the subject of so much attention?
Most likely, because it shares so many features of the numerically more
important Alzheimer's disease but is even more devastating to witness, and
because it can be transmitted through medical procedures that often involve
A poignant recent example of medically caused CJD is the outbreak of disease
in hypopituitary patients treated with native growth hormone that until 1985 was
extracted and processed from the Pituitary glands of cadavers. Some cadaver
donors had unsuspected CJD, and their glands were included in random batches of
hormone used to treat some 7000 plus patients in the United States. The
resulting contamination is responsible, to date, for 21 deaths, that, together
with the consequences of similar contaminations in England and France, account
for a total of nearly 100 deaths in treated adolescents and young adults, with
new cases continuing to occur after longer and longer incubation periods
following infections that occurred in the 1970's. Advances in biotechnology
supported by the NIH have made possible an unlimited supply of recombinant
growth hormone free of the risk inherent in the use of growth hormone of human
The growth hormone tragedy, and an even more recent outbreak of CJD in
neurosurgical patients who years ago had received contaminated dura mater grafts
- also from cadaver donors who had unsuspected CJD - has forced us to consider
with renewed concern any medical procedure that involves the transfer of tissue
or bodily fluids from one human being to another, and to try to predict where
further danger might be present and preventable.
This morning we are considering the possible risk that an individual
receiving blood (or a blood product) will contract CJD. The degree of this risk
depends upon a sequence of three probabilities: the probability that pooled
blood will contain a donation from at least one individual with CJD, the
probability that an individual receiving a therapeutic product made from such a
pool will be exposed to the infectious agent, and the probability that a
recipient who is so exposed will be infected and contract CJD.
The probability that pooled blood will contain a donation from a diseased
individual depends the prevalence of CJD in the donor population, and the number
of donors who contribute to the blood pool (donor pool size). The prevalence of
CJD in the United States is estimated to be one case per 1.3) million people,
and the size of donor pools in current practice typically ranges from a low of
10,000 donations to a high of 100,000 donations, although an occasional pool
reaches 400,000 donations. Using the 10,000 to 100,000 numbers, the probability
that a CJD patient will contribute to a pool of 10,000 donors is 0.8%; if the
pool size is increased to 100,000 donors, the probability rises to 7.6%.
Because it is unusual for a person who is already ill with CJD to donate
blood, our primary concern should be directed towards donations made during the
period before illness begins. Unfortunately, we do not know how long blood
might be infectious before a person shows the symptoms of CJD, but from studies
in experimental animals, we can make an educated guess that infectivity could be
present for as long as 10 years before the onset of symptoms. The prevalence
figure for a person "incubating" CJD would thus be 10-fold greater than the
prevalence of clinically apparent CJD, and the probability that a potentially
infectious individual would contribute to a donor pool becomes 7.6% for a pool
of 10,000 donors, and 55% for a pool of 100,000 donors.
The next step in the risk sequence - the probability that a recipient of
blood from a pool to which a CJD patient had contributed will be exposed to the
infectious agent - depends upon the amount of the infectious agent in the donor
pool, the number of particles of the agent needed to produce an infection, and
the number of recipients. Three different situations are possible.
First situation: the donor pool contains a large number of infectious
particles. For example, if a donor pool contains 10,000 infectious particles,
the pool is more or less "saturated": if given to a single recipient, one
infection will result- if divided among 100 recipients, it is extremely likely
that 100 infections will result, etc., until the number of recipients approaches
several thousand, when the increasing dilution and random distribution of
infectious particles will start to spare some recipients. This kind of situation
occurs in AIDS, where the blood of an HIV-infected donor may contain up to
100,000 infectious particles of virus.
Second situation: the donor pool contains a small number of infectious
particles. For example, if the pool contains 5 infectious particles, and the
donation is given to a single recipient, one infection will result; if divided
among 10 recipients, between one and five infections will result (a statistical
probability calculation). If the donation is divided among 100, or 1000, or
10,000 recipients, five infections will almost always result, as distribution
randomness will make it highly probable that each infectious particle goes to a
separate recipient. This is the most likely situation for CJD, based on
unpublished data from experimentally infected animals, but levels of infectivity
in the blood of human patients with CJD have never been determined.
Because we do not know with certainty what number of particles are needed to
produce an infectious unit, we must also consider the possibility that two or
more particles must join together to make a single infectious unit (for example,
if the required number were two, a specimen containing 100 particles would
contain 50 infectious units)
Third situation: one infectious unit consists of two interactive particles
that are independently distributed physical entities. For example, if a pool
containing 200 particles (100 infectious units) is given to 10 recipients, all
recipients are likely to be infected. If it is divided among 100, 1000, or
10,000 recipients, the random distribution of particles will result in fewer and
fewer individuals receiving the necessary two particles, and a progressively
decreasing number of infections will be observed, eventually reaching zero.
Thus, increasing the size of the donor pool would "dilute out" its infectivity.
The final step in the risk sequence - the probability that a recipient who
is exposed to the infectious agent of CJD will be infected - depends upon the
ease with which the agent can be transmitted by intravenous or intramuscular
administration. We do not know the answer to this question in humans, but in
experimental animals, these types of peripheral inoculations are 10 to 100 times
less effective than direct intracerebral inoculations in transmitting disease.
In summary, we conclude that although mathematical modeling gives us a
fairly precise idea of the influence of pool size upon risk of exposure,
estimates of the actual risk of a recipient contractina CJD from a contaminated
blood product are impaired by our lack of knowledge about most of the biological
factors that contribute to such estimates. Considering the magnitude of the
variables involved (low prevalence of CJD in the general population, low levels
of infectivity in the blood of CJD patients, and low efficiency of disease
transmission by intravenous or intramuscular administration), it seems likely
that the chance of contracting CJD from a pooled blood product to which a
patient with CJD has contributed is extremely small, no matter what the size of
the donor pool. The fact that epidemiological studies have so far been unable to
identify a single case of CJD resulting from the administration of blood or
blood products supports this contention.
Brown, P. Can Creutzfeldt-Jacob Disease Be Transmitted by Transfusion?
Current Opinion in Hematology 1995, 2:472-477
Abstract of manuscript
Summary of Future Research Proposals