. 19
( 45 .)


sistency m the incidence of sporadic CJD in most systematic surveys and there
is similar consistency in the sex distrtbutton, age distribution, and clinrcal fea-
tures. Any error m casetdenttfication must, therefore, be systematicand uniform.
A crucial assumption in CJD surveillance is that the dramatic nature of the
clrmcal illness will result m hospttalizatton and an accurate diagnosis of CJD.
Although current evidence suggests that there 1s a high degree of case
ascertainment in the young and middle aged, there is ajustifiable concern that
CJD may be missed in the elderly. The identification of previously undrag-
128 Will

Table 5
Diagnostic Criteria for CJD
1 Sporadic
a Defimte
1 Neuropathologrcally confirmed; and/or
n Immunocytochemrcally confirmed PrP posmve (Western blot), and/or
b. Probable:
1 Progressrve dementia
II Typical EEG
III At least two out of the followmg four clmrcal features myoclonus, visual
or cerebellar, pyramrdal/extrapyramrdal, or akmetrc mutrsm
c Possible:
1 Progressive dementia
11 Two out of four chmcal features listed above
in No EEG or atypical EEG
IV Duration <2 yr
2 Accidental transmrssron.
a Progressive cerebellar syndrome m a prturtary hormone recipient
b Sporadrc CJD with a recognized exposure risk
3 Familial.
a Definite or probable CJDplus defimte or probable CJD m a first degree relative
b Neuropsychratrrc disorderplus disease-specific PrP mutation

nosed CJD 1s exceptlonal in postmortem serves n-r the elderly demented and tt 1s
unlikely that large numbers of such cases are missed (23). In the UK study
there has been a stgntficant increase m the numbers of elderly patients with
CJD Identified m recent years (Fig. 3) but rt 1s of note that these elderly patients
have been identified from throughout the country and that the clinical features
in these cases were typical of sporadic CJD.
Recent studies of genetic forms of priori disease have established marked
phenotyptc vartation m the clnncal and pathological features of CJD (24) (and
GSS), and a stgnlficant proportton of these cases would not fulfil current diag-
nostic criteria for CJD. An important question is whether the failure to tdenttfy
thts type of case stgnrficantly preJudices the findings of epldemlologtcal sur-
veillance, and m turn this depends on an assessment of the incidence of atypt-
cal forms of CJD
Review of the clintcal features in pathologlcally confirmed or transmuted
cases of CJD indicates that approx 10% of cases of CJD are clnncally atypical
(25). Research in atypical genetic forms of CJD has depended on the tdenttfi-
cation of highly unusual pedigrees, some of which have been repeatedly stud-
Prion Disease Surveillance 129

Year 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94


Fig. 3. Sporadic CJD. Number of deathsin patientsagedover 75 yr (UK 198&1994).

ied (and published) over decades. The phenotype in genetic CJD varies with
the majority of cases readily identifiable as CJD, e.g., cases associated with a
codon 200 mutation (26) and some of the casesassociated with the codon 178
mutation. Only nine pedigrees of FFI have been identified worldwide (27) and
the estimated incidence of GSS is l/10 million/yr, a figure consistent with cur-
rent findings in the United Kingdom.
Mutations of the PrP gene have been sought in a range of neurodegenerative
disorders, almost uniformly without success(19,28-30). Recent evidence sug-
gests that progressive subcortical gliosis may be a prion disease (31,32) but
again this is an exceedingly rare condition. It has been suggested, on the basis
of isolated case reports, that prion disease may occur without characteristic
pathology (33) with the implication that all casesof dementia without diagnos-
tic pathology may be prion diseases(34). This is not supported by the failure of
transmission in laboratory experiments in which brain material from a variety
of cases of atypical dementia, including some with minimal pathology, were
inoculated into primates (I 9).
In conclusion, 100% case ascertainment cannot be achieved in CJD surveil-
lance. It is inevitable that some cases are missed, particularly in the elderly or
in relation to genetic forms of prion disease. However, current evidence sug-
gests that only small numbers of such casesare likely to be missed in view of
the rarity of genetic forms of CJD and that the epidemiological data on CJD is
therefore meaningful. The identification of novel human prion diseases, for
example in relation to BSE, may depend, by analogy with human growth hor-
130 Will

mone recipients and genetic forms of CJD, on systematic review of the
clinicopathologlcal phenotype as well as systematic study of the descrlptlve
eptdemlology of CJD

4. Case-Control Studies
The aim of the case-control study m CJD 1s to identify characteristics m the
patient group that are distinct from an age- and sex-matched control population,
thereby ldentlfymg a factor or factors that appear to increase the risk for the
development of CJD. Such studies have been carried out m the United States,
Israel, Japan, the United Kingdom and currently collaboratively m the EC.
There are many problems with this type of study m CJD The cause of spo-
radic CJD 1s unknown despite advances m basic science, and descriptive epl-
demlological studies have failed to provide any convincing evidence m relation
to the source of mfectivlty. Indeed, the apparently random distribution of cases
m space and time wlthm mdivldual countries and the worldwlde and consistent
incidence of CJD may be interpreted as indicating that an environmental risk
factor 1s unlikely.
Case-control studies in CJD are therefore not targeted but depend on an
assessment of a range of putative risk factors aimed at ldentlfymg case-to-case
transmission or cross-species transmission from animal prlon diseases. The
mcubatlon period m human prlon diseases may be extremely prolonged with a
mean incubation period of 13 yr m human growth hormone recipients (22) and
a range of mcubatlon periods from 5 to over 30 yr m kuru (I), mdlcatmg that
examination of risk factors must include analysis of potential exposure decades
m the past. This problem 1s compounded by the necessity of obtaining mfor-
mation from relatives of the patients. A maJor practical problem in case-con-
trol studies 1s the need to study CJD on a nationwide basis because of its rarity
and random occurrence
The first task in a case-control study 1s the formulation of a standard ques-
tionnaire addressing various potential areas of rusk. In the United Kingdom a
questionnaire has been constructed including sections on:
1. Past medical history (including prewous ocular or neurosurgery and blood dona-
tion/ reception);
2. Exposure to medlcmal products;
3 Residential history,
4 Occupational history;
5. Family history,
6 Dietary history; and
7 Exposure to animals
In order to avoid bias the questionnaire must be applied m a standard man-
ner, preferably by an individual experienced in utilizing the questionnaire. Cur-
Prion Disease Survedance 131
Table 6
Significant Risk Factors for CJD in Controlled Studies
Author Rusk factors
Bobowtck et al (36) 38 “selected” cases; None
healthy controls
Kondo and Kurorwa (3 7) Population study: Trauma in males
60 cases;
healthy controls
Kondo (38) 88 autopsred cases, Organ resection
autopsied controls
Davampour et al (39) Trauma or surgery to head or neck
26 cases,
40 controls Other trauma
Surgery needmg sutures
Davanipour et al (40) As above Roast pork, ham, underdone meat,
hot dogs
Davampour et al. (4f) As above Contact with fish, rabbits, squu-rels
Harrres-Jones et al. (42) 92 cases, Herpes zoster
184 controls Keeping cats
Contact with pets other than cats/
Dementia m family

rently m the United Kingdom and m the collaborative European study, a research
regrstrar visits each patrent m order to obtain clmical information and to intervrew
a relattve of the patient. A control case IS selected using the following crtterta.
1 Age match f 4 yr
2 Sex matched.
3. Inpatrent m the same hosprtal as the index case.
4. A relative of the same degree as the index case available for interview
5 Patient has a condmon clearly distinguishable from CJD
The first available control case fulfilling these criteria is selected for mter-
view, although m practtce it is unusual for more than one control case satisfy-
ing the criteria to be identified and, indeed, it IS not uncommon for no suitable
control to be available, requiring a further visit at a later date. The potential
difficulties of carrying out a case-control study in a rare disease on a nation-
wide basis should not be underestimated.
Alternative mechanisms for control selection have been apphed in CJD,
including the use of healthy relatives and random-digit dtallmg for community
controls, but regardless of methodology the results of these studies have been
largely negattve and no conststentrisk factor for CJD has been identified (Table 6).
132 Will

There are two maJor caveats m the mterpretation of the results of case-control
studies in CJD. The rarity of the disease inevttably results in wide confidence
intervals for any identified risk factor, particularly if this is an uncommon expo-
sure, whtch results m marked fragility of data from year to year. In the United
Kmgdom a number of statistically significant dietary rusk factors for CJD have
been identified since 1990 but these vary from year to year, mdtcatmg that any
apparent positive result should be treated with great caution Second, the rela-
tives of patients may be aware of the hypotheses being tested, a potential source
of bias that has become potenttally more important m view of the recent exten-
sive media coverage m relation to BSE and CJD. In the UK study, regular veal
consumption was identified as an apparently significant risk factor for CJD (odds
ratio: 13:32) m 1994. Analysis of the frequency of veal consumption m suspect
cases subsequently classified as “not CJD” demonstrated an almost identical
excessive exposure to veal, providmg powerful evidence of recall bias (6).
Although there are clear hmitations to the case-control methodology in CJD,
it is reasonable to conclude that currently there is no consistent evidence of a
specific environmental exposure that leads to increased risk of CJD. Imphcit m
this conclusion is that a number of potential risk factors, mcludmg blood trans-
fusion, previous surgery (including eye surgery), and exposure to animals can-
not be major risk factors for CJD. The accumulatton of evtdence m relation to
risk factors of CJD may, however, allow the identification of a sigmficant
change in relation to novel exposures.
5. Conclusions
The systematic study of the epidemiology of CJD 1slabor-intensive, time-
consummg, and costly. There are major practical difficulties m relation to case
ascertainment, vtsitmg hospitals on a nationwide basis, and case-control meth-
odology. The basic premise 1sthat casesof CJD are diagnosed and referred to
the surveillance center, which is dependent on a high level of cooperation from
the neurosctence community and others. In the United Kmgdom the geographi-
cal distributton of cases (Fig 4) indicates that cases have been consistently
identified m all regions of the Umted Kingdom and one measure of the level of
case ascertainment is the mctdence of CJD (Fig 5) One interpretation of the
rise m incidence IS that case ascertamment is improvmg (35) and there is clear
evidence in the United Kmgdom of an improvement m the identification of
CJD m the elderly. Co-operation with this type of study is dependent on mam-
taming good relations with referrmg physictans, a task that IS not made easier
by Intensive media scrutiny of individual cases.
The sea change in pubhc perception of priori diseases has occurred largely
because of the occurrence of BSE and despite the caveats regarding the sur-
veillance of CJD, it may only be by systematic epidemiological study that any
Prron Disease Surveillance

Ftg. 4. Geographical distrlbutlon of CJD in the United Kingdom. Definite and prob-
able cases (May 1, 1990-April 30, 1995)

risk to the human population from BSE may be identified. Should a significant
change m the pattern of CJD occur, it will be essential to consider whether this
IS related to the occurrence of BSE or some other factor. The incidence of CJD
has increased significantly from study period to study period m each country m
which serial surveillance of CJD has been carried out and DNA analysis has
resulted in an improvement m the identification of familial cases.It may only
be through comparisons of the epidemiology of CJD m different countries that
any significant change m one country can be identified.
The cause of CJD in any individual patient cannot be determined from epi-
demiological evidence. The occurrence of CJD in an adolescent in the United
Kingdom and the identification of three dairy farmers in the United Kingdom

0.9 -

0.7 -

0.6 -

0.5 -

0.4 -


02 -
01 '

Fig. 5. Incidence (per mrlllon) of CJD Defimte and sporadic cases(1970-1994)

with CJD has understandably led to concerns about potential links wtth BSE.
The avatlabthty of background eprdemlologrcal evtdence m other countrtes
may allow such atypical cases to be put mto perspective For example, three


. 19
( 45 .)