Nephrology Dialysis Transplantation 2007 22(Supplement 7):vii30-vii50; doi:10.1093/ndt/gfm328
© The Author [2007]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org
All patients receiving renal replacement therapy in the United Kingdom in 2005 (Chapter 4)
Ken Farrington1,
Raman Rao2,
Retha Stenkamp2,
David Ansell2 and
Terry Feest2
1Lister Hospital, Stevenage and 2UK Renal Registry, Bristol
Correspondence and offprint requests to: Ken Farrington, UK Renal Registry, Southmead Hospital, Southmead Rd, Bristol, BS10 5NB, UK. Email: ken.farrington{at}nhs.net
 |
Abstract
|
|---|
Summary data are provided for the whole United Kingdom. There
were 41 776 adult patients alive on renal replacement therapy
(RRT) in the UK at the end of 2005, a prevalence for adults
of 694 pmp. Addition of 748 children under the age of 18 on
RRT gives a total prevalence of 706 pmp. The more detailed analysis
includes data on 37 534 patients from 65 of the 70 units which
returned detailed data to the Registry: all in Northern Ireland,
Scotland and Wales, and 45 of the 50 units in England. The annual
increase in prevalence in the 38 renal units participating in
the Registry since 2000 was 5.0%. There is substantial variation
in the crude Local Authority area prevalence from 299 pmp to
1275 pmp.
In general, areas with large ethnic minority populations had high standardized prevalence ratios (SPR). Nevertheless several Local Authority areas in South Wales (Methyr Tydfil, Swansea and Rhondda/Cynon/Taff) had a higher SPR than would be predicted from the local ethnic mix. Another group in North West England (Bury, Rochdale, Oldham and Salford), had a lower SPR than expected from the local ethnic mix. The median age of prevalent patients on RRT was 56.6 years, that of patients on HD 64.5 years, PD 59.2 years and transplanted patients 49.7 years. The median vintage of the whole RRT population was 5.1 years: that of transplanted patients was 9.8 years, HD patients 2.8 years and PD patients 2.1 years. The maximal prevalence rate (SPR) occurred in men (2270 pmp) in the 75–79-year age band and women (1144 pmp) in the 65–74-year age band.
Of RRT patients in the UK, 45% had a transplant, 41.7% were on centre-based haemodialysis and 12% on peritoneal dialysis. The proportion of patients on home haemodialysis remained very small (1.2%) in spite of the recent NICE guidelines. The haemodialysis population is continuing to expand, mainly through growth in the proportion of patients undergoing dialysis in satellite units. The peritoneal dialysis population is continuing to contract in spite of the small but progressive rise in automated PD.
The most common identifiable diagnosis in those under 65 was glomerulonephritis (18.0%) and in those over 65 it was diabetes (13.4%).
One-year survival rates of prevalent patients in the different centres contributing to the UK Renal Registry are presented. The centres agreed to remove anonymity. There is no evidence of any significant differences in survival of prevalent patients between UK centres.
The one-year survival of prevalent dialysis patients increased significantly from 1998 to 2004 in England (83.3% to 87.1% P = 0.0001 for linear trend), Scotland (84.0% to 87.0% P = 0.023 for linear trend) and Wales (83.4% to 86.1% P = 0.027 for linear trend). The test for non-linearity in this trend (indicating that there has been a large increase which is now tailing off) was significant for England and Wales.
Keywords: chronic kidney disease; dialysis; end stage renal disease; epidemiology; haemodialysis; one-year survival; peritoneal dialysis; prevalence; quality improvement; renal replacement therapy; renal transplantation
|
Introduction
|
|---|
The prevalence data presented are from the whole United Kingdom.
In 2005, the UK Renal Registry received complete returns from
all five units in Wales, all five units in Northern Ireland
and 90% of the units in England. Data from all 10 units in Scotland
were obtained from the Scottish Renal Registry. In addition
summary data were obtained separately from the five remaining
English units not currently returning to the Registry, to enable
accurate calculation of prevalence and modality used.
Extrapolation from Registry data to derive other information relating to the whole UK was still necessary and these results must still be viewed with a little caution, although estimates become more reliable as coverage increases. The proportion of the population aged over 65 years was similar in the fully covered population (defined below, based on Local Authority (LA) areas whose population was thought to be fully covered by participating units) compared with the general population of England and Wales. The proportion from ethnic minority groups was lower in the fully covered population at 8.1% compared with 9.0% in the total population, because some areas not reporting to the Registry have catchments with high ethnic minority populations.
For comparisons between renal units and between local areas fully covered by the Renal Registry, the data from the Registry are fully valid. Data on children and young adults can be found in Chapter 13.
|
All adult patients receiving renal replacement therapy in the United Kingdom (31 December 2005)
|
|---|
There were 41 776 adult patients receiving RRT in the United
Kingdom at the end of 2005, giving a total population prevalence
for adults of 694 pmp (
Table 4.1). Addition of the 748 children
under age 18 on RRT (Chapter 13) gives a total prevalence of
706 pmp.
In those renal units continuously reporting for the last 6 years
there was an average rise in prevalence from year to year of
between 4.2% and 6.5%.
|
Prevalent patients by renal unit on 31 December 2005
|
|---|
For 2005, detailed data on prevalent patients were returned
from 45 of the 50 renal units in England, all five units in
Wales, all five units in Northern Ireland and all 10 units in
Scotland, a total of 37 534 patients. The number of prevalent
patients in each renal unit and the distribution of their treatment
modalities are shown in
Table 4.2.
There is a wide variation in the number of prevalent patients
in each unit and in the distribution of these patients in the
different treatment modality categories. This is due to many
factors including geography, local population density, age distribution,
ethnic composition and social deprivation index of that population.
Local facilities and preferences also play a role in determining
the modality distribution. Some of these will be discussed later
in the chapter. However, another major factor is whether or
not the renal unit is also a transplant centre. The 23 renal
units which are also transplant centres tend to have a higher
proportion of transplant patients under follow up compared with
the other 42 units, but are also the larger dialysis units.
The transplant/dialysis ratio is markedly higher in transplant
centres than in other renal units (1.17
vs 0.46:
P < 0.001).
The wide variability of this ratio both in transplanting (0.58–2.65)
and non-transplanting (0.01–1.2) renal units suggests
considerable variation in policies for follow up of transplanted
patients; some transplant centres continue to follow up the
patients they transplant for other renal units, others transfer
them back to their parent unit but at variable times post transplant
and some renal units do not follow up any transplant patients.
|
Changes in prevalence 2000–2005
|
|---|
The total number of prevalent patients in all 65 centres contributing
to the Registry in 2005 is 41 776. The increase from 2004 to
2005 in the 59 centres with data in both years was 4.6%, which
is entirely consistent with 2000–2005 analysis. For individual
centres, the changes in total numbers are shown in
Table 4.3.
There were wide variations between centres with respect to change
in prevalent patient numbers between 2004 and 2005, ranging
from an 18.6% increase (Clwyd) to a 5.5% decrease (Airdrie).
In some units (Wrexham, Hammersmith and Charing Cross, Leicester
and Oxford) changes in the prevalent population are partly due
to changes in catchment areas. This explanation is confirmed
by the fact that the prevalence changes for the local authority
areas served by these units have been consistent with national
trends.
The growth in prevalent patient numbers in the UK since 1982 is shown in Figure 4.1.
The total percentage increase in number of prevalent patients
in the 38 renal units who have returned data continuously from
2000 to 2005 was 27.8%. The rate of increase was similar in
England (27.6%), Scotland (28.6%) and Wales (27.9%) and fairly
uniform over the time span, varying between 4.2 and 6.5% per
year (
Table 4.4).
|
Local Authority prevalence
|
|---|
The prevalence of RRT and standardized prevalence ratios in
those Local Authorities with complete coverage in 2005 are shown
in
Table 4.5.
View this table:
[in this window]
[in a new window]
|
Table 4.5. Prevalence of RRT and standardized prevalence ratios in local authorities with complete coverage by the Registry
|
|
|
Standardized prevalence ratios
|
|---|
Methods
The methods of calculating the standardized rate ratio are described
in detail in Appendix D (
www.renalreg.org). In summary, age-and
gender-specific prevalences were first calculated using the
available registry data on the number of prevalent patients
for the covered area in England, Wales, Scotland and Northern
Ireland and the data on the age and gender breakdown of the
population of each Local Authority area obtained from the 2001
census data from the Office of National Statistics (ONS). These
age and gender prevalences were then used to calculate the expected
prevalence for each LA area. The age and gender standardised
ratio is therefore equal to (observed prevalence)/(expected
prevalence).
A ratio of 1 indicates that the LA area's prevalence was as expected if the age/gender rates found in the total covered population applied to the LA area's population structure; a level above 1 indicates that the observed prevalence is greater than expected given the LA area's population structure; if the lower confidence limit was above 1 this is statistically significant at the 5% level. The converse applies to standardised prevalence rate ratios under one.
Prevalence estimates of RRT in relatively small populations such as those covered by individual Primary Care Trusts incur wide confidence intervals for any observed frequency. To enable assessment of whether an observed prevalence rate differs significantly from the national average, Figures 4.2 and 4.3 have been included. For any size of population (X axis), the upper and lower 95% confidence limits (dotted lines) around the national average prevalence can be read from the Y axis. Any observed prevalence for renal failure outside these limits is significantly different from the national average. Thus for a population of 50 000, an observed prevalence outside the limits of 470–930 pmp is significantly different, whilst for a population of 500 000 the limits are 625–770 pmp.
Results
There were substantial variations in the crude LA area prevalence
from 299 (Bury) to 1275 pmp (Carrickfergus). As discussed earlier,
local authorities with small populations have wide confidence
limits for standardised prevalence rate (SPR), such that the
interpretation of an individual year may be difficult. Nevertheless
the annual standardized prevalence rate is inherently more stable
than the annual standardized acceptance.
Geographical considerations and ethnicity are the major factors underlying the variation in SPR. There were 33 local authority areas with a significantly low SPR, 123 with a normal SPR and 51 with a significantly high SPR. The geographical distribution of these is summarised in Table 4.6. The North West (P < 0.0001) and the South East of England (P = 0.03) had a significantly higher proportion of areas with a low SPR, whilst in London, Wales, Scotland and Northern Ireland, the proportion was significantly lower (P = 0.03 in all cases). Conversely, London (P < 0.0001) and Northern Ireland (P = 0.048) had a significantly higher proportion of areas with a high SPR, whilst in the North East (P = 0.04) and the North West of England (P = 0.008), the proportion was significantly lower. Although overall areas with a high SPR had significantly higher ethnic minority populations than areas with significantly low or normal SPRs (P < 0.0001) (Figure 4.4), in some areas such as South Wales, ethnicity does not seem to be a major factor.
View this table:
[in this window]
[in a new window]
|
Table 4.6. Summary regional distribution of local authority areas with significantly low, normal, or significantly high values of SPR and mean% non-White
|
|
The relationship between the ethnic composition of a LA area
and its SPR is further demonstrated in
Figure 4.5, which shows
the relationship between ethnicity and SPR for all local authorities
with available data. A small group of local authority areas
in Wales have a higher SPR than might be predicted from the
local ethnic mix. These are Methyr Tydfil, Swansea and Rhondda/Cynon/Taff.
Another small group of local authority areas in the North West
of England, have a lower SPR than might be expected by the local
ethnic mix. These are Bury, Rochdale, Oldham and Salford. It
is unlikely that social deprivation alone can account for these
disparities. Further investigation would be of interest. Tower
Hamlets appears to have an inappropriately low SPR for what
is the second highest proportion of non-Whites in the Registry.
|
Vintage of prevalent patients
|
|---|
Table 4.7 shows the median vintage (years since starting renal
replacement therapy) of prevalent RRT patients in 2005. Median
vintage of the whole RRT population was 5.1 years. Patients
with functioning transplants had survived a median 9.8 years
on RRT whilst the median vintage of HD and PD patients was much
less (2.8 and 2.1 years, respectively). The dialysis population
is of course much older and would be expected to have shorter
survival. This is not a substantial change from the 2004 data.
|
Age of prevalent patients
|
|---|
The median age of prevalent patients on RRT was 56.6 years (
Table 4.8).
The age profile is markedly different in patients on dialysis
than in transplanted patients. The median age of patients on
HD (64.5 years) was higher than that of patients on PD (59.2
years) and substantially higher than that of transplanted patients
(49.7 years). There were wide variations in median age between
renal units for the whole RRT population (50.8–67.7 years).
The major determinant of the median age of the prevalent RRT
population is the ratio of the number of transplant and dialysis
patients in that population (
r = –0.764,
P < 0.0001).
The differing age distributions of transplant and dialysis patients
are well illustrated in
Figure 4.6, the maximum prevalence of
dialysis patients being almost two decades later than transplant
patients. In patients under the age of 65 years, 56.3% of prevalent
RRT patients had been transplanted with 43.7% on dialysis. The
proportions were dramatically different in older patients, with
21.2% having been transplanted and 78.8% on dialysis.
|
Gender
|
|---|
In the United Kingdom there were more patients in the age range
55–65 years than in any other decade in both males and
females (
Figure 4.7). However the ‘corrected’ peak
prevalence, expressed as SPR calculated from local authority
populations covered by the Registry using 2001 Census data,
occurred in the age band 65–74 (1565 pmp) overall, but
was different in men (peak 75–79-year age band; 2270 pmp)
from women (peak 65–74-year age band; 1144 pmp:
Figure 4.8).
Furthermore the male:female ratio of prevalence increased markedly
with age from 1.48 in the 25–34 age band to 4.46 in those
greater than 85 years.
|
Ethnicity
|
|---|
Thirty-six of the 65 centres submitting data to the Registry
provided ethnicity data that were at least 90% complete. The
data for centres with less than 50% returns for ethnicity are
excluded from
Table 4.9. Centres in Scotland are not required
to report ethnicity to the Scottish Registry.
|
Primary renal disease
|
|---|
In the previous 2 years’ reports, the statement with the
help of table indicating that diagnosis code GN histologically
not examined (EDTA code 10) had been included in the ‘Uncertain’
group for analysis, was incorrect. Approximately 1000 patients
had been incorrectly allocated to the glomerulonephritis category.
Table 4.10 this year, has now been corrected. The previous years’
data has also been retrospectively analysed to this grouping
and the data this year show no change and is consistent with
the reports prior to 2004.
The most common specific diagnosis overall remains glomerulonephritis,
in contrast to the pattern in incident patients in whom diabetes
predominates. This reflects different survival and different
ages of the patients with these diagnoses.
There are age-related differences. The prevalence of the aetiology uncertain/glomerulonephritis—not biopsy proven category is much greater in those aged over 65 years (27.7% vs 19.8%). In addition, diabetes (13.4%) [not glomerulonephritis (9.9%)] was the most common specific diagnosis in those over 65 years. The male:female ratio was significantly greater than unity for most primary renal diseases, but only marginally for polycystic kidney disease and pyelonephritis. The ratio for polycystic kidney disease is similar to that in incident patients and the possible underlying reasons were discussed in Chapter 3. The ratio for pyelonephritis is markedly different in prevalent (1.1) and incident patients (1.7). This is a consistent finding and may indicate poorer survival on RRT of males with this diagnosis.
The distribution of patients between the modalities is also heavily influenced by primary renal diagnosis (Table 4.11). Patients with pyelonephritis, polycystic kidney disease and glomerulonephritis are much more likely to have been transplanted than patients with diabetes and those with renal vascular disease. The differences are even more marked in patients over the age of 65.
View this table:
[in this window]
[in a new window]
|
Table 4.11. Transplant:dialysis ratios by age and primary renal disease in the prevalent RRT population (31 December 2005)
|
|
|
Diabetes
|
|---|
The median age of all prevalent diabetic RRT patients (58.8
years) is slightly higher than that of non-diabetics (56.2 years),
patients with Type 1 disease being considerably younger (52.8
years) than those with Type 2 disease (66.6 years) (
Table 4.12).
The RRT vintage of prevalent diabetics both Type 1 (3.3 years)
and Type 2 (2.2 years) is significantly less than that of prevalent
non-diabetics (5.8 years). Fewer prevalent diabetics than non-diabetics
have transplants (26.9%
vs 48.2%): 36.1% of patients with Type
1 disease and only 10.3% of those with Type 2 disease. The proportions
are even lower in patients over the age of 65 (
Table 4.13).
View this table:
[in this window]
[in a new window]
|
Table 4.12. Type of diabetes, median age, gender ratio and treatment modality in prevalent RRT patients (31 December 2005)
|
|
|
Modalities of treatment
|
|---|
The most common treatment modality is transplantation (45.0%),
closely followed by the proportion on centre-based HD (41.7%)
as shown in
Figure 4.9. The proportion of patients on home HD
remains very small (1.2% of RRT) in spite of the recent NICE
guidelines.
View larger version (17K):
[in this window]
[in a new window]
[Download PowerPoint slide]
|
Fig. 4.9. Treatment modality in prevalent RRT patients 2005. In some centres local coding of RRT modality is such that the Registry could not differentiate between CAPD and cycling PD. In these centres all PD patients are included as CAPD disconnect. Thus the proportion of PD patients on cycling PD is a slight underestimate.
|
|
Transplantation is the predominant treatment modality in patients
less than 65 years old and haemodialysis in those 65 or older
(
Table 4.14). The proportions are similar in all of the UK countries
except a small preference in favour of HD over PD in Northern
Ireland, particularly in older patients.
Haemodialysis is increasingly prominent with increasing age
at the expense of transplantation. The proportion of each age
group treated by PD remains fairly stable across the whole age
spectrum (
Figure 4.10).
|
Haemodialysis
|
|---|
The proportion of dialysis patients on HD in the UK was 78%
and higher in those over 65 years old than in younger patients
(83%
vs 74%). The proportions varied widely between renal units
but the same pattern of age distribution was maintained in all
but five units (Dorset, Ulster, Inverness, Dumfries & Galloway
and Wolverhampton,
Figure 4.11). A slightly larger percentage
of the male dialysis population (78.7%) were on HD than of the
female dialysis population (76.7%:
P < 0.001).
The proportion receiving HD in satellite units varied. Twenty-nine
units had no satellite haemodialysis whilst 12 units dialysed
more than 50% of their haemodialysis patients in satellites
(
Figure 4.12). Satellite HD amounted to 34.5% of total HD activity.
Twenty-one units had no home HD programme. In the 44 units which
did offer home HD, the proportion of HD patients treated by
this modality ranged from 0.6% to 11.1%. Overall only 2.7% were
on home HD. Twelve units had home HD programmes amounting to
more than 5% of total HD activity.
|
Peritoneal dialysis
|
|---|
The proportion of prevalent dialysis patients on PD varies widely
ranging from 2.4% (one patient) in Ulster to 38.2% in Ipswich
(
Figure 4.13). Overall 23.3% of the female dialysis population
were on PD compared with 21.2% of the male dialysis population
(
P < 0.001). The overall male to female ratio was 1.4 but
there was marked variation between centres, the ratio varying
from 0.6 to 5.0.
CAPD using disconnect systems remains the most common PD mode
(62.0% of all patients on PD). The use of automated PD (APD)
is continuing to increase and now comprises 32.2% of all PD
treatments. However, the use of APD varies widely between units,
ranging from 0% to 100% of all PD treatments (
Figure 4.14).
Treatment for six or more nights weekly is the norm, but many
units use less frequent treatments on an occasional basis and
one unit (Guys), exclusively. Use of connect systems remains
very uncommon (3.6% of all treatments).
|
Change in treatment modality 1997–2005
|
|---|
Top
Abstract
Introduction
