The National CJD Surveillance Unit

Western General Hospital

Edinburgh EH4 2XU

and

Dept of Infectious and Tropical Diseases

London School of Hygiene and Tropical Medicine

Keppel Street, London WC1E 7HT

C O N T E N T S

1. Summary

The national surveillance programme for Creutzfeldt-Jakob disease (CJD) in the UK was initiated in May 1990. In 1999, the National CJD Surveillance Unit (NCJDSU) became a WHO Collaborative Centre for Reference and Research on the Surveillance and Epidemiology of Human Transmissible Spongiform Encephalopathies (TSEs). The information provided in this ninth report continues to provide evidence of a high level of case ascertainment. Detailed clinical and epidemiological information has been obtained for the great majority of patients. A high post mortem rate has been maintained through the period of the study 1990-2000. The success of the project continues to depend on the extraordinary level of co-operation from the neuroscience community and other medical and paramedical staff throughout the UK. We are particularly grateful to the relatives of patients for their help with this study.

The average number of cases of sporadic CJD identified annually since 1990 was higher than in previous surveillance periods extending back to 1970. It is impossible to say with certainty to what extent these changes reflect an improvement in case ascertainment and to what extent, if any, they reflect changes in incidence.

In 1990-2000 mortality rates from sporadic CJD in England, Scotland, Wales and Northern Ireland were, respectively, 0.75, 0.86, 1.00 and 0.46/million/year. The difference between the rates in each country is not statistically significant (p=0.3). These rates are comparable to those observed in other countries in Europe and elsewhere in the world, including countries which are free of BSE. There was some variation in the observed mortality rates between the different regions within the UK but this variation is not statistically significant (p=0.5). The highest and lowest mortality rates from sporadic CJD were observed in the East Anglia (SMR=129) and East Midlands regions of England (SMR=81). Previous analyses have found no convincing evidence of space-time clustering, and this remains the case for the analyses in this report.

Up until 31 December 2000, there have been 84 deaths from definite or probable variant CJD (vCJD) in the UK (in addition 2 probable cases died in January 2001 and a further 7 probable cases remained alive as at 31 January 2001). Of the 84 deaths to 31 December 2000, 75 were confirmed neuropathologically with a further two awaiting neuropathological confirmation. The clinical, neuropathological and epidemiological features of all these cases of vCJD are remarkably uniform and consistent with previous descriptions. However, a case of neuropathologically confirmed vCJD in an individual who died aged 74 years in October 1999 significantly extends the age range in vCJD.

Analysis of the incidence of vCJD by standard region suggests that the incidence of vCJD in the "North" of Great Britain may be higher than in the "South". The rate ratio (north vs south) based on place of residence in 1991 was 1.81 (95% CI 1.20, 2.74). The mean Carstairs' deprivation score for areas of residence of people with vCJD was -0.41 (95% CI -1.02, 0.19), which is close to the national average of zero. Regional rates of vCJD were correlated with consumption of other meat or meat products as classified and recorded by the Household Food Consumption and Expenditure Survey (r =0.73), but not with similar data from the Dietary and Nutritional Survey of British Adults. Five people with vCJD in Leicestershire formed a cluster (p=0.004).

Risk factors for the development of vCJD include age, residence in the UK and methionine homozygosity at codon 129 of the prion protein gene - 87 cases of vCJD with available genetic analysis have all been methionine homozygotes. The analyses in this report do not provide conclusive evidence of an increased risk of vCJD associated with past surgery, previous blood transfusion, occupation or a range of dietary factors. However, the power of the case-control study, from which these results are derived, is limited by the relatively small number of cases and controls. For some putative risk factors, such as blood transfusion or surgery, it may be many years before an accurate assessment of risk can be made because of the likely prolonged incubation periods.

The national surveillance of CJD was initiated in May 1990 in response to a recommendation in the Report of the Working Party on Bovine Spongiform Encephalopathy (Southwood Committee). The surveillance is funded by the Department of Health and by the Scottish Executive Health Department. The initial aim of the NCJDSU was to identify any change in the pattern of CJD that might be attributable to the emergence of bovine spongiform encephalopathy (BSE). Such a change was recognised in 1996 when vCJD was first described. The NCJDSU now aims to monitor the characteristics of all forms of CJD, to identify trends in incidence rates and to study risk factors for the development of disease. This report documents the findings from the NCJDSU (UK) in relation to cases of sporadic, familial, iatrogenic and vCJD diagnosed up to 31^st December 2000 (with data ascertained up to 31^st January 2001).

2.1 Sporadic Creutzfeldt-Jakob disease

Between 1st January 1970 and 31st December 2000, 888 cases of sporadic CJD were identified in the UK, of which 7 cases were still alive on 31st December. One case was identified in Jersey, which was not included in the following UK analyses. Of these cases, 689 (78%) were classified as definite cases with the remainder classed as probable. Figure 1a shows the number of deaths each year from sporadic CJD for England and Wales between 1970 and 2000, Figure 1b shows similar data for Scotland and Northern Ireland between 1985 and 2000 and Figure 1c shows the combined number of deaths (i.e. UK) from 1985 to 2000. In England and Wales the number of deaths identified each year has increased from an average of about 10 per year at the beginning of the 1970s, to about 40 per year in the 1990s. A similar phenomenon has been observed in other European countries and this probably largely reflects improved case ascertainment. Over the shorter time period for which data are available for Scotland and Northern Ireland there is no clear secular trend although the highest numbers of cases are seen in the two most recent years for which data are complete. Over the period 1990-2000 the average annual mortality rates from sporadic CJD per million population were 0.75 in England, 1.00 in Wales, 0.86 in Scotland and 0.46 in Northern Ireland, as shown in Table 1. When account is taken of age and sex, the variation in recorded mortality between the different countries is not statistically significant (p > 0.2) (unchanged).

Figure 1a Deaths from sporadic CJD, England and Wales, 1970-2000

Figure 1b Deaths from sporadic CJD, Scotland and Northern Ireland, 1985-2000

(please note different scale from Figure 1a)

Figure 1c Deaths from sporadic CJD, UK, 1985-2000

Note: Data for 2000 are not yet complete.

Table 1 Deaths from definite and probable sporadic CJD by region and county of death: 01/05/1990 - 31/12/2000

Figure 6 shows the observed numbers of deaths by quarter with the fitted underlying trend and 95% confidence interval.

Figure 6

Observed (-o-) quarterly incidence of vCJD deaths.

Fitted underlying trend (__) is given with its 95% confidence limits (...)

The fitted line is log(incidence) = -0.092 + 0.075*quarter. This gives an estimated annual increase of 1.35, 95% CI (1.13 to 1.61) and a p-value for the trend of < 0.001.

The estimate of the current quarterly incidence of deaths is 7.4. Note that it is possible there may be more deaths in the latest quarter which have not yet been reported. The large number of deaths in the fourth quarter of 98 means that the variation by quarter is more than would be expected by chance, this is accounted for by re-scaling the variance in the model when assessing the significance of the trend.

From the model, if the current trend continues, the predicted total number of deaths for 2001 is 36 with a 95% confidence interval of 21 to 58.

Another analysis was performed in which the incidence of onsets and delay to diagnosis was jointly modelled using Poisson regression. This analysis included 88 cases diagnosed by the end of 2000 and showed an increasing trend in onsets from 1994 to 2000 similar to that found for deaths (Figure 7). The analysis also demonstrated a significant reduction in the delay from onset to diagnosis with an estimated reduction from 14.5 months for onsets in 1995 to 9.5 months for onsets in 2000. The model also estimated that approximately a further 23 onsets will have occurred by the end of 2000 which were not diagnosed by the end of 2000.

 Figure 7 Observed (-o-) and expected (-e-) quarterly incidence of vCJD onsets

Table 3 Deaths from definite and probable vCJD by region and county of onset: 1 May 1995 - 31 January 2001

Table 4 and Figure 9 show cumulative regional rates of vCJD based on cases' place of residence in 1991 and the population aged 10 years and above at that time. We previously performed an analysis of the first 51 cases, distinguishing two areas. The "North" comprised four standard regions: Scotland, North, Yorkshire and Humberside, North West. The "South" comprised the remaining 6 regions: Wales, West Midlands, East Midlands, East Anglia, South West, South East.

Table 5 shows the distribution of cases between the "North" and the "South" in 1991, distinguishing between those cases included in the previous analysis and those classified as cases subsequently (under an a priori hypothesis). The excess of cases previously identified in the "North" (rate ratio controlling for age and sex = 1.94; 95% c.i. 1.12, 3.36) appears to have been largely maintained in subsequent cases (rate ratio controlling for age and sex = 1.66; 95% c.i. 0.89, 3.09). Taking all 91 cases together the estimated rate ratio controlling for age and sex is 1.81 (95% c.i. 1.20, 2.74).

Table 5

Comparison of cumulative incidence in the "North" of the UK (excluding Northern Ireland) with that in the "South"

*North versus South, adjusted for age and sex.

Northern cases were slightly older at onset than southern cases (median of 27 years versus 24 years; p=0.25) and more of them were male (60% versus 48% of southern cases; p = 0.24).

To investigate whether the difference between North and South might be related to socio-economic differences, we examined the Carstairs' index for each case based on their place of residence (enumeration district [ED]) in 1991. The mean Carstairs' score for the cases was -0.41 (95% c.i. -1.02, 0.19), close to the national average (0.0), and to the average conditional on their regional distribution (0.11). The cases were evenly distributed across the five quintiles of the index (data not shown).

Data from the Dietary and Nutritional Survey of British Adults on consumption of those items most relevant to putative transmission of BSE through diet are presented in Table 6. Those items most likely to have contained mechanically recovered meat (MRM) or high titre material from the central nervous system ("burgers and kebabs", "sausages", "meat pies and pastries", "other meat products") show no consistent pattern of higher consumption in the northern regions. While consumption of "meat pies and pastries" appears to be highest in the north, consumption of "burgers and kebabs" and "other meat products" appears highest in the South East.

Table 6 Regional variations in weekly quantities of foods consumed (grams) measured in the period 1986-1987^*

* Data from the Nutritional Survey of British Adults

The data on meat consumption published in the Household Food Consumption and Expenditure report for 1988 for the period 1984-1986 are shown in Table 7. Only four categories of meat/meat products were distinguished. "Carcase meat" includes joints, steaks, chops and mince of beef, pork and lamb. The products most likely to have contained bovine MRM or high titre material from the central nervous system fall into the "other meat and meat products" category. There is some evidence of a positive correlation between consumption of "other meat and meat products" and vCJD incidence (Spearman's rank correlation = 0.73; p=0.02) (Figure 10). Interestingly both carcase meat and poultry consumption show some evidence of a negative correlation with vCJD incidence (Spearman's rank correlation coefficients of -0.65 (p=0.06) and -0.62 (p=0.08) respectively), while bacon consumption shows no evidence of a correlation with vCJD incidence (Spearman's rank correlation coefficient = 0.18; p=0.64).

Table 7

Regional variations in weekly quantities of foods brought into the home (grams) measured in the period 1984-1986^*

*Data from Household Food Consumption and Expenditure: 1988

The spatial scan statistic applied to data at the census enumeration ward level identified a group of five cases in Leicestershire as the "most likely cluster" (P=0.004). No other statistically significant (P<0.05) clusters of cases were identified. In particular, the cases in Kent, which have been the subject of debate, were not identified as a cluster. The total population of Leicestershire at the 1991 census was approximately 870,000, giving a cumulative incidence of vCJD for Leicestershire of about 5.7 per million, compared with an overall cumulative incidence in the UK (excluding Northern Ireland) of 1.7 per million. Four of the five Leicestershire cases lived in the district of Charnwood, with a population in 1991 of about 142,000, giving a cumulative incidence of vCJD for this district of about 28.2 per million. The fifth case lived a couple of km outside Charnwood District and the greatest distance between any two of the cases in 1991 was less than 10km. In other respects, these cases do not appear, at first sight, different from cases from elsewhere in the country. Two were female and the ages at onset ranged from 17 to 33 years. All five cases were reported to have eaten beef products. One had worked as a farm labourer. A summary of the final report of the investigation into the North Leceistershire cluster of vCJD undertaken by Dr Gerry Bryant and Dr Philip Monk can be found on the

Since 1970, up to 31st December 2000, 42 cases of CJD attributable to iatrogenic exposure have been identified, 6 in individuals receiving dura mater implants, 35 in individuals who had received human-derived growth hormone (hGH) and one in a recipient of human gonadotrophin (hGN). The mean age at death of the hGH /hGN group was 29 years (with a range of 20-45 years) and for the dura mater cases 43 years (range 27-59 years).

The first identified iatrogenic case was a dura mater recipient who died in 1979. The first hGH-related death occurred in 1985.

Forty cases of familial CJD have been identified since 1970, excluding cases of GSS. Of these, 38 were resident in England and 2 were resident in Wales. Fourteen of the cases had insertions in the coding region of the PrP gene, 11 carried the mutation at codon 200 (Glu-Lys), 2 at codon 178 (Asp-Asn, both with methionine at codon 129, ie FFI), 1 at codon 117 (Ala-Val) and 1 at codon 210 (Val-Ile). Eleven were identified as familial on the basis of relatives known to have had CJD. The mean age at death was 55 years (with a range 38 - 68 years).

The distribution of codon 129 genotypes in sporadic CJD has been analysed since the inception of the Unit in 1990. The overall distribution of codon 129 genotypes in sporadic CJD (69% MM, 14% MV, 17% VV) (see Table 8) is consistent with findings from other European countries. There is no evidence (p > 0.1) of a change in the codon 129 distribution in sporadic CJD between the periods 1990-1995 and 1996-2000.

 Table 8 Codon 129 genotypes of cases of sporadic CJD in the UK, 1990-2000

All cases for whom genetic data are available (87) were methionine homozygotes at codon 129 of the PrP gene.

A laboratory dedicated to the analysis of cerebrospinal fluid (CSF) was established in the CJDSU in March 2000. The principal role of this laboratory is the analysis of CSF for 14-3-3 and related brain-specific proteins in suspect cases of CJD. The laboratory provides a diagnostic service for the United Kingdom. In addition, the laboratory has received samples from other countries in relation to the European collaborative surveillance projects with which the Unit is involved. Finally, the laboratory is involved in research into CSF proteins in relation to CJD.

In the period from March to December 2000, a total of 201 CSF samples were received and processed by the laboratory. The samples are divided into 3 groups. Firstly, there are samples from patients formally referred to the NCJDSU as suspect CJD cases. Secondly, there is a group of "CSF only referrals". These are CSF samples analysed by the laboratory from patients in the United Kingdom with illnesses that were thought not to be CJD and thus not formally referred as suspect cases. Finally, there is a group of samples from other countries. The details are found in the Table 9.

Table 9

Number and origin of CSF samples received and analysed at the NCJDSU during March 2000 - December 2000

Table 10 shows the CSF 14-3-3 results in these groups of patients.

Table 10

CSF 14-3-3 results in NCJDSU UK suspect referrals (n=99), March-December 2000

CSF 14-3-3 was detected in 91% (20/22) of samples from patients with definite and probable sporadic CJD.

One patient with definite sporadic CJD was negative for 14-3-3 in the CSF. The neuropathological features of this case were typical of those found in sporadic CJD who are valine homozygous at codon 129 of the PRNP gene.

A second patient with definite sporadic CJD had 2 samples. The first sample, which was taken early in the course of the disease was negative. The second sample taken 3 months later, was positive. This case was heterozygous (MV) at codon 129 of the PRNP gene and had a relatively long duration illness of 13 months.

Of the 11 probable sporadic cases (classified as probable based on clinical features plus positive 14-3-3), 3 also showed the typical EEG appearances of sporadic CJD (generalised triphasic periodic complexes at approximately one per second). In the remaining 8 cases, EEG did not show the typical EEG appearances of sporadic CJD.

Three patients referred as suspect CJD who turned out to have another illness, were found to be positive for 14-3-3. One of these patients was suffering from paraneoplastic syndrome, the second from an encephalitis, while the third patient had no certain diagnosis. However, this last patient's clinical condition improved, excluding the possibility of sporadic CJD.

CSF 14-3-3 was detected in 67% (4/6) of patients with definite variant CJD. However, in the 9 cases classified as probable variant CJD, none had a positive 14-3-3. This may be explained by the timing of the CSF sample.

CSF 14-3-3 was not detected in any of the patients initially referred as suspect variant CJD who later turned out to have another illness.

Of the 72 CSF samples from non-suspect cases, 5 were blood stained and therefore unsuitable for analysis, 3 were positive for 14-3-3 and the remainder did not have detectable 14-3-3. Of the 3 positive 14-3-3 cases, one had Herpes Simplex encephalitis and 2 had encephalitis of unknown cause.

In conclusion, the analysis of CSF 14-3-3 has a high degree of sensitivity and specificity for the diagnosis of sporadic CJD, and has enabled patients who would have been previously diagnosed as possible CJD to be classified as probable. In particular, a diagnosis of probable sporadic CJD was made in 8 patients (during the period March-December 2000) who would not have fulfilled the current WHO and EU criteria due to either the absence of typical EEG changes or because an EEG had not been performed.

The role of the co-ordinator is to provide advice on all forms of CJD to the patient, their family and professional carers, including information on the clinical features, diagnostic procedures and prognosis. The co-ordinator assists with the co-ordination of the care locally, working closely with and supporting the identified key worker who has responsibility for ensuring that the needs of the patient and the family are met. Regular case conferences are encouraged, which the co-ordinator attends when possible in order to assist with the assessment and reassessment of evolving needs. Having experience with families previously affected by this disease allows the co-ordinator to provide information on successful interventions, which may be helpful for other families.

When a referral has been made to the unit of a suspect case of CJD, the co-ordinator makes direct contact with the family and arranges to meet with them within two weeks. Thereafter the co-ordinator meets with the patient and the family on a regular basis, depending on need, to provide support and is available by telephone between visits. Support continues after the patient dies.

Since the establishment of this post in June 2000, the co-ordinator has had contact with 20 families affected by CJD or vCJD and has visited each on a regular basis. Currently there are nine patients that are affected by vCJD and the co-ordinator has contact with each family.

The Transfusion Medicine Epidemiology Review (TMER) is a collaborative project between the UK NCJDSU and UK Blood Services (UKBS). The main purpose is to investigate whether there is any evidence that CJD or vCJD may have been transmitted via the blood supply.

vCJD cases (definite and probables) are notified to the UKBS by NCJDSU; a search establishes whether any have acted as donors. Donation records are checked and all components traced through hospital records. Details of recipients are forwarded to NCJDSU for subsequent checking.

 In the reverse procedure, patients with vCJD reported to have received blood transfusions are identified by NCJDSU and notified to UKBS. Details of transfusions are traced through hospital records and relevant blood donors identified. The identity of donors is notified to NCJDSU for subsequent checking.

For vCJD cases, 15 were reported to have been blood donors. To date, only 10 have been traced at blood centres, and 8 of these had donated blood, with a resulting 48 blood components. It has been established that 22 components were transfused to named recipients.

In the reverse study, 8 vCJD cases were reported to have received blood transfusions. Checks revealed that 2 were not transfused, 2 had transfusions which predated available records and 4 had records of transfusion which could be traced. To date, these 4 received 117 blood components of blood, corresponding to 111 named donors (one patient received 103 components). The donors of two components are not traceable and checking is still incomplete on 4 components.

No donors or recipients identified in the study through the tracing of donation and transfusion records appear in the NCJDSU register as cases. Further data on vCJD cases, to define time interval between blood donation and development of disease are being accumulated.

(Collaborators on this project: Dr P.E. Hewitt and Dr C.A. Llewelyn).

The aim of this project is to use the mechanism of the British Paediatric Surveillance Unit to identify all cases of progressive intellectual and neurological deterioration in children in the UK, particularly those with features suggestive of vCJD. All cases are discussed by an expert neurological advisory group of six paediatric neurologists which allocates the cases to a diagnostic category. (Verity CM, Nicoll A, Will RG, Devereux G, Stellitano L. Variant Creutzfeldt-Jakob disease in UK children: a national surveillance study. Lancet 2000; 356: 1224-1227).

After 45 months surveillance, 1082 patients with suspected PIND have been reported. Among them were four cases of vCJD; three definite and one probable. Three were reported in 1999 and one in 2000. One girl was aged 12 at onset - the youngest ever case of vCJD. No other children with the clinical features of vCJD were identified. The expert group has discussed 785 cases, of which 435 have a confirmed underlying cause other than vCJD, being categorised into 89 known neurodegenerative diseases.

(Collaborators: Dr C. Verity, Dr A. Nicoll, Ms G. Devereux).

3. Case-Control Study

Following LREC approval, the complex and time-consuming process of control recruitment commenced. As of March 2001, letters have been sent to 79 GPs to ask them to participate in the study and forty-four practices have been visited. Letters have been sent to 723 potential controls by their GP describing the study and asking them to agree to be approached directly by the NCJDSU. One hundred and seventy- two (24%) of these individuals have replied, of whom 127 (74%) agreed to be contacted by the NCJDSU. Of those written to by the NCJDSU (n=122), 88 (72%) have replied. Eighty-seven have consented to take part in the study and have nominated a relative for interview. Sixty-nine relatives have agreed to take part in the study, of whom 67 have been interviewed.

By the end of 2000, risk factor questionnaires for 53 community controls and 60 hospital controls had been completed (relating to 25 and 60 vCJD cases, respectively). Of the vCJD cases included in the analyses, 18 have both community and hospital controls, seven have only community controls and 42 have only hospital controls.

Comparisons of cases with controls have been performed separately for the community and hospital controls, because of the potential heterogeneity of these two control groups with respect to factors under investigation.

Analysis was restricted to cases with matched community controls because of the potential for bias in using hospital controls to investigate medical risk factors.

Seventy-two percent of community controls were reported to have had some sort of operation/surgical procedure in the past compared with 64% of cases (Table 11). There is no evidence to suggest that particular operations/surgical procedures are associated with increased risk of vCJD (Table 11). However, the confidence intervals around these odds ratio estimates are wide. These findings should not be interpreted as evidence that transmission via surgery has never, or could never, occur.

Table 11

Reported operations/surgical procedures for 25 cases of variant Creutzfeldt-Jakob disease and 53 community controls.

*Includes appendicectomy

Reported consumption of different types of meat from 25 cases¹ of variant Creutzfeldt-Jakob disease and 53 community controls.