[1]
Date: Wed 10 Jan 2001 20:12:16 +1100
From: Robert Flower <rflower@med.usyd.edu.au>
The Epidemiology of Variant CJD: Some Additional Questions
----------------------------------------------------------
In common with Nick Honhold, the author of the initial posting, I have no
particular expertise in Creutzfeldt-Jakob disease (CJD), or bovine
spongiform encephalopathy (BSE), but as a blood banker I have a very real
interest. I, too, found an epidemiological approach taking a rolling
average to smooth annual fluctuations in the figures [recently posted in
CJD (NEW VAR.) - UK: UPDATE 3 JAN 2001] of interest.
A 4-year rolling average in the data presented for GSS (fifth column from
the date) shows:
1985 - 1988: one death
1989 - 1992: one death
1993 - 1996: 12 deaths
1997 - 2000: 2 deaths (no cases in 1999 or 2000)
Whatever statistical test one applies, the peak between 1993 and 1996 is
very significant. The fact that the number of new cases has fallen to zero
for 1999-2000 argues that this peak was not due to increased diagnostic
vigilance.
The question that this observation raises for me is, "Is it possible that
exposure to the variant CJD prion in individuals genetically susceptible to
GSS results in development of GSS rather than variant CJD?" A range of
mutations associated with susceptibility to CJD (and GSS) has been reported
in the literature and, for iatrogenic CJD, disease develops in genetically
susceptible individuals following exposure to prions via preparations of
growth hormone or gonadotrophin. If this hypothetical form of the disease
(GSS-initiated as a result of exposure to the variant CJD prion) is a real
phenomenon, and the peak between 1993 and 1998 is a result of exposure of
individuals genetically susceptible to GSS to the variant CJD prion, this
would provide evidence that for some groups (individuals with GSS
mutations) disease resulting from of exposure to the variant CJD prion in
the UK has already peaked.
For 1999-2000 no deaths were reported, a contrast with the peak 2-year
period 1995-1996, in which 7 deaths were reported. This would also suggest
that (in at least some cases) the disease process following exposure to
variant CJD is defined by the genetic make-up of the individual (exposed to
the variant CJD prion) rather than the disease process being defined as a
result of exposure to an "infectious dose" of the variant CJD prion per se.
Is there any evidence that individuals with variant CJD have a susceptible
phenotype, and thus develop the particular form of the disease diagnosed as
variant CJD? Is it possible that in some individuals with a genetically
defined susceptible phenotype, the pathogenesis of the disease initiated
following exposure to the variant CJD prion is that of "classical" CJD? Are
data in relation to the PrP phenotype of the "sporadic" and "variant CJD"
cases publicly available?
A final question stems from an extension of the logic of this argument to
resistance to disease. It may be of value to ask: "What is the basis of
resistance to variant CJD (and prion diseases)?", in addition to the
question "What is the basis of susceptibility?" A range of mutations
associated with susceptibility to CJD (and GSS) has been reported, but is
it possible that some individuals inherit factors producing a resistant
(rather than a susceptible) phenotype?
For example, a particularly active cell-surface chaperonin may refold the
variant CJD prion after the individual is exposed, thus destroying its
infectivity in that "resistant" individual. If resistance is based on a
process of this type, it is possible to hypothesize that these individuals
do not carry an infectious form of the prion despite exposure to variant CJD.
--
Professor RLP Flower
PaLMS Transfusion Service &
Department of Haematology and Transfusion Medicine
Northern Sydney Health & University of Sydney
Royal North Shore Hospital
Pacific Hwy
St LEONARDS NSW 2065
Australia
<rflower@med.usyd.edu.au>
******
[2]
Date: Wed 10 Jan 2001 11:04:26 +0000
From: O. Hotz de Baar <Oz@upthorpe.demon.co.uk>
Analysis of Outbreak Data by Date of Onset
------------------------------------------
The initiator of this thread [Nick Honhold <nick@plewland.demon.co.uk>]
posed the following question: "In most diseases, date of onset is used for
generating epidemic curves. But, in general, variant CJD is being reported
by date of death. Would there be any change in the data if it is looked at
by date of onset"?
[Here is a partial answer]. Unfortunately an untimely lightning strike
caused the loss of some of my data and results. However I still have a
simple curve fit for the dates of *onset* for the UK variant CJD [cases]
which gave the following likely curves. Please note that [the] data [are]
only given to monthly resolution. Also please note the considerable
potential delay between onset and the reporting of the date of onset, which
could exceed 12 months.
Exponential form:
Cumulative cases (onset) = A + exp(KD)
A couple of options are the following:
#1 / #2
A = offset = -35 / -25 days
K = constant = .00053 / .00063
D = days from = Jun 75 / Jul 79
Notes:
1) Nothing should be read into D as it effectively incorporates a constant
of proportionality.
2) As with all exponential predictions the best fit is critically affected
by new cases being reported in the fitted period.
3) Predictions going into the future will rapidly deviate and going any
significant time forward should be considered untrustworthy.
4) There is no apparent deviation that implies a peak is being approached.
A marginally better fit can be had using a bilinear fit; e.g.
Cumulative cases = .0264 (days from Dec 93) + .0276 (days from July 97)
(Obviously one ignores negative values here).
Notes:
1) A possible reasons for the bilinear form might be two infection events
with an approximate linear case incident rate.
2) Another (as suggested above) might be a change in reporting about July 97.
3) I found the similarity of the two slopes striking when I first got the
results.
For convenience I copy a list of the Hansard onset dates from 1 Jan 1994
(first case) accurate to the start of the reported month to within a few days.
0,31,59,151,181,212,334,334,365,365,365,424,546,577,608,638,699,699,730,730,790,790,790,851,851,1004,1035,1065,1065,1127,1155,1277,1277,1277,1308,1369,1369,1400,1400,1400,1430,1430,1430,1461,1492,1492,1520,1551,1581,1581,1581,1612,1612,1642,1642,1642,1704,1704,1734,1826,1826,1857,1885,1916,1916,1916,1916,1946,1946,1977,1977,2007,2038,2099,2130,2130,
2160,2191.
It's probably not worth looking at any new figures until late spring 2001.
--
O. Hotz de Baar
<Oz@upthorpe.demon.co.uk>
******
[3]
Date: 10 Jan 2001 07:37:10 -0500 (EST)
From: Dr. Georg Hoffmann <GHoffmann@Trillium.de>
The Reliability of Diagnostic Tests and Other Questions
-------------------------------------------------------
The diagnosis of variant versus sporadic CJD is based mainly on a specific
(?) histological picture, the so-called florid plaque. What is missing is
the proof of presence or absence of bovine prions in human brain, blood,
fluids, etc. (No specific and sensitive test is available so far). There
are other non-specific differences like the age of onset (50 years often
being taken as a threshold for sporadic CJD) or the absence of a certain
pattern in the EEG. Established criteria such as sensitivity and
specificity, predictive value, etc. have not been applied so far to these
diagnostic tests (as far as I know), and are probably hard to apply at all
because of the low number of cases. Altogether the differentiation seems to
be more than weak.
It is interesting to note that in Britain every third CJD case occurring
under the age of 50 has been identified as variant CJD, whereas in Germany
none has been identified among a total of 75 comparable cases! This is to
be seen in the light of the officially reported bovine spongiform
encephalopathy (BSE) epidemiology in Germany. The figure was zero despite
[several] reported cases in close neighbors (France, Switzerland) prior to
24 Nov 2000, when an unfortunate farmer took a voluntary test and ruined
himself overnight due to the positive result.
An interesting epidemiologic question is the (increased?) susceptibility of
carriers of the homozygous M/M variant in the PRNP gene. It has been
reported that all British patients tested so far showed this genotype,
which normally has a prevalence of only 39%. Interestingly enough, M/M
carriers among the South Fore cannibalistic tribes were the first to suffer
from the Kuru endemia, which has many similarities with variant CJD except
that the infectious prions were of human rather than bovine origin (J
Infect. Dis. 183, 192, 2001).
--
Prof. Dr. med. Georg Hoffmann
Editor of Laboratory Management News
Hauptstr. 12 b
D-82284 Grafrath, Germany
<GHoffmann@Trillium.de>
[More interesting questions and a partial answer. It is an unfortunate
circumstance that many of these questions may not be amenable to
epidemiological analysis until the number of cases increases or hopefully
declines dramatically. - Mod.CP]
.
A ProMED-mail post
http://www.promedmail.org
ProMED-mail, a program of the
International Society for Infectious Diseases
http://www.isid.org