New adult patients starting renal replacement therapy in the UK in 2005 (Chapter 3)
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 |
|---|
 Top Abstract IntroductionAdult patients accepted for...Local changes in acceptance...Geographical variation in...EthnicityAgeGenderPrimary renal diagnosisFirst established treatment...Survival of incident patientsLate referral of incident... |
|---|
In 2005, the acceptance rate for renal replacement therapy (RRT) in adults in the UK was 108 per million population (pmp). This was derived from complete data for adults in the UK, as data were obtained separately from the five English renal units not currently returning to the Registry. In addition, 87 children started RRT (see Chapter 13) giving a total incidence of 110 pmp.
From 2001 to 2005 there has been an 7.3% rise in the acceptance numbers in those 42 renal units with full reporting throughout that period.
In the UK, for adults in 2005, the crude acceptance rates in Local Authorities (LA) varied from 0 (in two very small LA areas in Scotland and Northern Ireland) to 271 pmp; the standardized rate ratios for acceptance varied from 0 to 2.76. Excluding the two areas with null returns, 20 areas had significantly low ratios, all of them in England. Thirty had significantly high ratios, seven in Northern Ireland, four in Scotland, three in Wales and seven in London.
Over the period 2001–2005, 25 areas had a significantly low standardized acceptance rate; 24 in England and one in Scotland. All except one of these had ethnic minority populations of <10%. Thirty-seven had high standardized acceptance rates, seven in Scotland where ethnicity data were not available, 14 from areas with ethnic minority populations in excess of 10%, and 12 were in Wales or the Southwest of England.
The median age of patients starting RRT in England has increased from 63.8 years in 1998 to 65.2 years in 2005. The median age of incident non-White patients is significantly lower at 56.8 years.
In England, the acceptance rate is highest in the 75–79 age band at 408 pmp, as in Scotland at 580 pmp; in Wales the peak is in the 80–84 age band at 525 pmp, as in Northern Ireland with a rate of 825 pmp.
Diabetic renal disease (20%) remains the most common specific primary renal disease. There was a significant positive correlation between the percentage of incident RRT patients with diabetic renal disease and the percentage of non-Whites in the incident cohort.
Haemodialysis (HD) was the first modality of RRT in 76% of patients, peritoneal dialysis (PD) in 21% and pre-emptive transplant in 3%. In 1998, the proportion whose first modality was HD was 58% and this continues to increase.
By day 90, 8% had died, a further 1% had stopped treatment or been transferred out leaving 91% of the original cohort on RRT. Of these, 71% were on HD, 26% on PD and 3% had received a transplant.
Data on first referral to a nephrologist were available from 22 centres for the period 2000–2005 (for a total of 5611 patients and 59 centre-years).
In 2005, the mean percentage of patients referred late (<90 days before dialysis initiation) was 30% (centre range 13–48%). This was similar to the value in 2000.
Patients referred late were older, a higher proportion of them were male, a lower proportion non-White, and a lower proportion with no recorded comorbidity. Patients with polycystic kidney disease and diabetic nephropathy tended to be referred early compared with the whole incident cohort and those with uncertain aetiology and no recorded diagnosis referred late.
Estimated GFR (eGFR) at the start of RRT appears to be higher in older than younger patients. eGFR is significantly lower in those referred late compared with those referred earlier and this is especially marked in the older patients.
The geometric mean eGFR of all patients starting RRT rose from 6 in 1997 to above 7.5 in 2003, since when it has remained stable.
|
Introduction |
|---|
|
TopAbstractIntroductionAdult patients accepted for...Local changes in acceptance...Geographical variation in...EthnicityAgeGenderPrimary renal diagnosisFirst established treatment...Survival of incident patientsLate referral of incident... |
|---|
The acceptance data presented were from the whole UK. In 2005, the UK Renal Registry received complete returns from all five renal units in Wales, all five renal units in Northern Ireland and 90% of the renal units in England. Data from all 10 renal units in Scotland were obtained from the Scottish Renal Registry. In addition, summary data were obtained separately from the five remaining English renal units not currently returning to the Registry, to enable accurate calculation of acceptance rates and initial 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 subsequently, based on Local Authority (LA) areas whose population was thought to be fully covered by participating renal 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.
|
Adult patients accepted for renal replacement therapy in the UK, 2005 |
|---|
|
TopAbstractIntroductionAdult patients accepted for...Local changes in acceptance...Geographical variation in...EthnicityAgeGenderPrimary renal diagnosisFirst established treatment...Survival of incident patientsLate referral of incident... |
|---|
Overall take-on rate
In 2005, there were 6485 adult patients who started renal replacement therapy (RRT) in the whole UK. This equates to acceptance rates of 108 per million population (pmp) for adults (Table 3.1) and 110 pmp including children. This represents an overall increase in the past 2 years. The adult acceptance rate in England was 104 pmp. Acceptance rates in Wales, Scotland and Northern Ireland were all higher than this, at 129, 122 and 140 pmp, respectively (Figure 3.1). There continues to be very marked gender differences in the take-on rate, the annual acceptance was 137 (95% CI 132–141) pmp in males and 81 (95% CI 77–84) pmp in females.
|
|
|
Local changes in acceptance rate |
|---|
|
TopAbstractIntroductionAdult patients accepted for...Local changes in acceptance...Geographical variation in...EthnicityAgeGenderPrimary renal diagnosisFirst established treatment...Survival of incident patientsLate referral of incident... |
|---|
Acceptance rates of individual units
The number of patients accepted by each renal unit is shown in Table 3.2. There is variation in time trends between renal units, which may reflect chance fluctuation, completeness of reporting, changing incidence of established renal failure (ERF), changes in referral patterns or catchment populations and areas and the introduction of conservative care teams.
|
The percentage change over the period 2001–2005 is shown for those 42 renal units, which had full reporting during that period and for the same data on a national level. Overall there has been an 7.3% rise in the acceptance numbers. There are wide variations between different renal units, the more extreme ones are related to changes in catchment populations, e.g. an increase of 70% since 2001 (Hull), a decrease of 25.8% (Liverpool). The Northamptonshire region has moved from the Oxford renal unit catchment to that of the Leicester renal unit. The increase seen in the national figures is similar to that reported for the period 2000–2004 in last year's report. Acceptance rates of individual renal units have not been calculated, as their catchment populations are not precisely defined.
|
Geographical variation in acceptance rates in England, Northern Ireland, Scotland and Wales |
|---|
|
TopAbstractIntroductionAdult patients accepted for...Local changes in acceptance...Geographical variation in...EthnicityAgeGenderPrimary renal diagnosisFirst established treatment...Survival of incident patientsLate referral of incident... |
|---|
Introduction
Equity of access to RRT is an important goal of service provision. The need for RRT depends on social and demographic factors including age, gender, social deprivation and ethnicity, so comparison of crude acceptance rates by geographical area alone can be misleading. This section, as in previous reports, uses age and gender standardization and ethnic minority profile to compare RRT incident rates. The impact of social deprivation was recorded in the 2002 report. The population used for standardization is the sum of all LA areas for which the Registry had full coverage in 2005.
Methods
Standardized acceptance rate ratios were calculated as detailed in web Appendix D (www.renalreg.org). Briefly, age- and gender-specific acceptance numbers were first calculated using the available registry data on the number of incident patients for the covered areas of England, Wales, Scotland and Northern Ireland. The age and gender breakdown of the population of each LA area was obtained from the 2001 Census data from the Office for National Statistics (ONS), and used to calculate the expected age- and gender- specific acceptance numbers for each LA area. The age- and gender- standardized acceptance rate ratio is the observed acceptance numbers divided by the expected acceptance numbers. A ratio below 1 indicates that the observed rate is less than expected given the LA area's population structure. This is statistically significant at the 5% level if the upper confidence limit is less than 1.
Results
Local Authority acceptance rates
Acceptance rates in LAs with complete coverage by the Registry are shown in Table 3.3.
|
Acceptance rates for RRT in relatively small populations such as those covered by individual Primary Care Trusts or LAs have wide confidence intervals for any observed frequency. To enable assessment of whether an observed acceptance rate differs significantly from the national average, Figure 3.2 has been included.
|
For any population size (X-axis), the upper and lower 95% confidence intervals around the national average acceptance rate (dotted lines) can be read from the Y-axis. (The example plot shown in Figure 3.2 assumes that the national average is 109 pmp). An observed acceptance rate outside these limits is significantly different from the national average. In order to be judged as significantly different from national norms the observed take-on rate for a population of 50 000 would have to be outside the limits of 17–200 pmp/year, whilst for a population of 1 million, the limits are from 88 to 129 pmp/year.
In the UK, for adults in 2005, the crude acceptance rates in LAs varied from 0 [in two very small LA areas; in Scotland (Eilean Siar—population 26 502) and Northern Ireland (Moyle—population 15 932)] to 271 pmp. There were also wide variations in the standardized rate ratios for acceptance (0–2.76). Excluding the two null returns alluded to above, 20 areas had significantly low ratios, all of them in England. Four of these had ratios <0.5: Salford (0.35), Darlington (0.37), Isle of Wight (0.46) and Poole (0.47). Thirty had significantly high ratios: 7 in Northern Ireland, 4 in Scotland, 3 in Wales and 7 in London. Nine had ratios of 2.0 or more. Six of these were in Northern Ireland [Antrim (2.58), Armagh (2.00), Carrickfergus (2.73), Castlereigh (2.50), Coleraine (2.66) and Cookstown (2.76)] one in Scotland [Dundee City (2.20)] and the others in London [Newham (2.10) and Greenwich (2.11)].
In Table 3.3, the trends over the 5 years, 2001–2005 are shown, illustrating the wide variations in annual standardized acceptance ratios in areas with small populations, especially those with habitually low take-on rates.
Also depicted in Table 3.3 are the standardized acceptance ratios derived from combined 2001–2005 data. Only data from areas with 3 or more years’ data are included in the following analysis. This excludes data from Northern Ireland because data were only available for 1 year. Twenty-five areas had a significantly low take on rate (shaded and italicized in Table 3.3), 24 in England. All of these had ethnic minority populations of <10% (except Oldham, 13.9%). Nine areas had a standardized acceptance ratio less than 0.7 (excluding two Scottish areas with very small numbers). These were Isle of Wight, Blackpool, Stirling, Hertfordshire, Wiltshire, Lancashire, Bury, Salford and Oldham. Thirty-seven had significantly high standardized acceptance ratios (shaded and bold in Table 3.3). Seven of these were in Scotland and ethnicity data were not available, and 14 had ethnic minority populations of >10%. Of the remaining 16, 12 were in Wales or the Southwest of England.
In Figure 3.3, standardized acceptance ratios derived from these combined data are plotted against the percentage of non Whites in the general population (ONS 2001 census) corresponding to the same area. It can be seen that in general, areas with a high ethnic minority population (and/or a socially deprived population, as shown in previous reports) have high standardized acceptance rate ratios; although some areas with a very low ethnic minority population also have high standardized acceptance rate ratios. These age standardized rates (Table 3.3) are all relative to an overall acceptance rate which still needs to be adjusted for social deprivation and ethnicity so that the population RRT requirement can be calculated.
|
|
Ethnicity |
|---|
|
TopAbstractIntroductionAdult patients accepted for...Local changes in acceptance...Geographical variation in...EthnicityAgeGenderPrimary renal diagnosisFirst established treatment...Survival of incident patientsLate referral of incident... |
|---|
Only 30 of the 65 renal units which submitted returns (46%) provided 90% or more complete ethnicity data (Table 3.4 includes only centres with 50% or more returns). Nevertheless, this is an improvement on previous years. The percentage of renal units providing ethnicity data <50% complete also improved (i.e. decreased) to 31% (20 units). This degree of incompleteness still makes analysis of ethnicity data unreliable.
|
Within the renal units with over 90% returns there is a huge variation in the percentages of new patients from the ethnic minorities ranging from 0% (Belfast, Antrim, Newry, Tyrone, Ulster, York, Gloucester, Carlisle, Airdrie) to over 40% (Royal Free, Bradford, Hammersmith and Charing Cross and Barts/Royal London). The latter renal units all include areas with high standardized acceptance rates.
|
Age |
|---|
|
TopAbstractIntroductionAdult patients accepted for...Local changes in acceptance...Geographical variation in...EthnicityAgeGenderPrimary renal diagnosisFirst established treatment...Survival of incident patientsLate referral of incident... |
|---|
The median ages of patients starting RRT are 65.2 years in England, 68.3 years in Northern Ireland, 65.4 years in Scotland, 67.5 years in Wales and 65.5 years for the whole UK (Table 3.5). Within the UK, there was a small increase in the median age of patients starting RRT from 63.9 years in 1998 to a plateau of 65.5 years in 2002.
|
In England, the acceptance rate is highest in the 75–79 age band at 408 pmp, as in Scotland at 580 pmp; in Wales the peak is in the 80–84 age band at 525 pmp, as in Northern Ireland with a rate of 825 pmp (Table 3.6).
|
The median age of incident UK non-White patients in 2005 was considerably lower, at 56.8 years, than that of whole incident cohort (P < 0.001; Wilcoxon test). This probably reflects the lower median age of the ethnic minority populations compared with the White population.
There remain large variations by centre in median age of new patients (Figure 3.4), the maximum (Tyrone) and the minimum (Barts and the London) are separated by over 2 decades. There are many possible reasons for these differences relating to local population demographics and the proportion of ethnic minorities in the catchment area. There may be differences in the prevalence, nature and management of renal disease and in approaches to conservative management.
|
|
Gender |
|---|
|
TopAbstractIntroductionAdult patients accepted for...Local changes in acceptance...Geographical variation in...EthnicityAgeGenderPrimary renal diagnosisFirst established treatment...Survival of incident patientsLate referral of incident... |
|---|
As in previous years there was an excess of males starting RRT in all age groups (Figure 3.5). The ratio of males to females is fairly constant until the age of 75, but males are increasingly represented in older patients (Figure 3.6).
|
|
The mean UK male to female ratio in the 2005 incident cohort is 1.6 : 1. All reporting centres except Gloucester, Dumfries and Galloway and Dunfermline report an excess of males in the 2005 incident cohort (Figure 3.7). The renal unit male to female ratio varies from 0.94 (Gloucester) to 6.5 (Carlisle). These high ratios are likely to be an effect of small numbers. All five renal units with a male to female ratio >2.5 in 2005 had a total take on number of 35 or less in that year.
|
|
Primary renal diagnosis |
|---|
|
TopAbstractIntroductionAdult patients accepted for...Local changes in acceptance...Geographical variation in...EthnicityAgeGenderPrimary renal diagnosisFirst established treatment...Survival of incident patientsLate referral of incident... |
|---|
The distribution of new patients by age, gender and cause of ERF is shown in Tables 3.7 and 3.8. For most types of kidney disease the male to female ratio is >1.5 : 1, as expected. The exception is adult polycystic kidney disease (APKD) for which the ratio approaches 1, as in the 2004 report. This would be expected from the mode of inheritance. Patients with APKD are relatively young when they develop ERF; approximately 4 times as many commence RRT in the under 65 cohort than the older cohort. This contrasts with renal vascular disease which is over 5 times more common in the older cohort. The gender imbalance may relate in part to the presence of factors, such as hypertension, atheroma and renal vascular disease, which are more common in males, and more common at increasing age. These factors may influence the rate of progression of renal failure.
|
|
The proportion of null returns for primary renal diagnosis has increased from a UK mean of 9.2% in 2004 to 12.0% in 2005. There is considerable national variation from 8.9% in Wales, through 10.1% in England, 16.9% in Northern Ireland, to 25% in Scotland. There is also very marked variation between centres (Table 3.8).
As in previous cohorts the diagnosis of aetiology uncertain/glomerulonephritis unproven is the most common and in patients over the age of 65 accounts for 
30% of all diagnoses. Some centre variation with respect to this diagnosis is likely to reflect the lack of clear definition of certain diagnostic categories e.g. hypertensive disease and renal vascular disease; some may result from differences between centres in the degree of certainty required to record other diagnoses. In keeping with this there are significant negative correlations between the frequency of the aetiology uncertain diagnosis and those of diabetes, glomerulonephritis, pyelonephritis and renal vascular disease.
Diabetic renal disease remains the most common specific primary renal diagnosis in the UK, at about 20%. Diabetic kidney disease generally follows the pattern of population distribution of ethnic minorities, but is also related to social deprivation. In the 33 centres with >70% ethnicity returns, and excluding four centres who classified 
60% of their patients as having an uncertain diagnosis, there was a significant correlation between the percentage of incident RRT patients with diabetic renal disease and the percentage non-Whites in the incident cohort (r = 0.60, P < 0.001). Five of the eight centres (62.5%) with 20% non-Whites in their incident cohort had a mean incidence of diabetic renal disease in that cohort of >25%, compared with only 1 of 25 (4%) centres with less non-Whites (P = 0.001: Fisher's exact test)
|
First established treatment modality |
|---|
|
TopAbstractIntroductionAdult patients accepted for...Local changes in acceptance...Geographical variation in...EthnicityAgeGenderPrimary renal diagnosisFirst established treatment...Survival of incident patientsLate referral of incident... |
|---|
In the UK, in 2005, haemodialysis (HD) was the first modality of RRT in 75.5% of patients, peritoneal dialysis (PD) in 21.4% and pre-emptive transplant in 3.1% (defined as first treatment recorded irrespective of any later change). This represents little change from the figures recorded in the 2004 report but a significant change from 1998 when the very first treatment modality was HD in 57.7%. Many patients, especially those referred late to renal units, undergo a brief period of HD before being established on PD. As an indication of the elective treatment modality, the established modality at 90 days is more representative. By day 90 of treatment, 8.4% had died, a further 1.2% had stopped treatment or been transferred out, leaving 90.4% of the original cohort on RRT. Of these remaining patients, 70.6% were on HD, 26.2% on PD and 3.2% had received a transplant (Figure 3.8).
|
In Table 3.9, these variables are represented as a percentage of the whole 2005 cohort, showing for the whole UK, 63.8% on HD, 23.7% on PD and 2.9% with a transplant. The percentage of the incident cohort which had died by day 90 varied considerably between individual renal units (0–35%). Small numbers are the likeliest explanation for these differences. Both of the two renal units with zero death rate and six of the seven units with a death rate above 15% took on less than 45 patients during the year. In addition, the median age of incident patients was >68 years in six of the seven with the higher death rate.
|
There were major differences between individual renal units in the percentage of new patients established on HD at 90 days (range 38–100%, Figure 3.9). Only two renal units had <50% on HD, whilst 19 had
80%. A significantly higher proportion (P < 0.0001) of incident dialysis patients over the age of 65 (82.0%) were on HD at 90 days compared with their younger counterparts (63.7%) (Table 3.10). This translates to the proportion of patients on PD being twice as high in patients aged <65 years as the proportion in older patients (36.3 vs 18.0%). This trend appears to be increasing. These overall differences were reflected in the vast majority of renal units though in five the proportions were reversed and PD was more popular in the elderly (Dorset, Ulster, Clwyd, Inverness and Southend). The male : female ratio in patients on HD was 1.70 compared with a ratio of 1.57 for patients on PD.
|
|
|
Survival of incident patients |
|---|
|
TopAbstractIntroductionAdult patients accepted for...Local changes in acceptance...Geographical variation in...EthnicityAgeGenderPrimary renal diagnosisFirst established treatment...Survival of incident patientsLate referral of incident... |
|---|
This analysis is to be found in Chapter 12.
|
Late referral of incident patients |
|---|
|
TopAbstractIntroductionAdult patients accepted for...Local changes in acceptance...Geographical variation in...EthnicityAgeGenderPrimary renal diagnosisFirst established treatment...Survival of incident patientsLate referral of incident... |
|---|
Methodology
Data were included from all incident patients in the years 2000–2005 with the following exceptions:
- All patients under 18 years of age at the start of RRT.
- All Scottish data since the date first seen by a nephrologist is only available for a handful of people.
- The small number of patients who recovered sufficient renal function to allow discontinuation of dialysis.
Referral time was calculated as the number of days between the date of first being seen by a nephrologist and the date of RRT initiation. A small proportion of data (1.8%) was excluded because of actual or potential inconsistencies. Only data from those centres/years with 75% completeness were used. Centre/years where 10% of the referral times were zero were excluded. After these exclusions, data on 5611 patients were available for analysis. Referral times of 90 days or more were defined as early referrals. Referral times of less than 90 days were defined as late referrals. Twenty-nine people were calculated to have negative referral times (–1 to –14 days). These were attributed as zero. After the exclusions outlined, the data available for analysis are detailed in Table 3.11, which shows the percentage completeness of data from the centres and years included in the data set.
|
Late referral by centre and year
The percentage of patients referred to a nephrologist less than 90 days before RRT initiation in the included centres and years in the period 2000–2005 is shown in Table 3.12. The range in 2005 was 13–48%. The mean annual incidence of late referral in 2005 was 29.8%, which was similar to the value in 2000.
|
Time referred before dialysis initiation in the 2005 incident cohort
Just over half the patients (52.6%) had been referred over a year before they needed to start dialysis. There were 10.3% of patients referred within 6–12 months, 7.3% within 3–6 months and 29.8% within 3 months.
Age and late referral
Patients who were referred late (<90 days before dialysis initiation) were significantly older than patients referred earlier (median age 67.7 vs 64.3 years; P < 0.001). Furthermore, the median duration of pre-dialysis care diminished progressively with increasing age beyond the 45–54 age group (Figure 3.10).
|
Gender and late referral
There was a borderline significant difference in the male : female ratio in those referred late (<90 days) and those referred earlier (1.79 vs 1.59; P = 0.047), with late referral more common in males.
Ethnicity, social deprivation and late referral
Patients from the Chinese ethnic minority were excluded from this analysis as the numbers with referral data were too small (n = 17). Thirty patients with an ethnic background of ‘other’ were also excluded. The proportion of non-Whites (South Asian and Black) referred late (<90 days) was significantly lower than in Whites (21.7 vs 27.7%; P = 0.012), implying that late referral may be less common in non-Whites. This will be partly due to the high incidence of diabetes in non-Whites (which tends to be referred earlier) and the Whites being an older group. Advancing age is also associated with late referral.
Importantly in the UK, there was no relationship between social deprivation and referral pattern.
Primary renal disease and late referral in 2005 incident cohort
Late referral (<3 months) differs significantly between primary renal diagnoses (Table 3.13, 
2 test P < 0.001). Multiple comparison tests between the different diagnoses groups have not been made as there would be a high risk of producing a significant test by chance. Patients with a diagnosis of ‘other identified category’ or ‘not sent’ appear to have higher rates of late referral, those with diabetes and polycystic disease have lower rates.
|
Modality and late referral
Referral pattern had a marked effect on initial modality choice. The proportion of patients whose initial modality was PD was significantly less in the late referral group in comparison to the group referred earlier (13.2 vs 31.8%; P < 0.0001). By 90 days after dialysis initiation the difference was partially redressed, though the proportion on PD was still significantly lower after late referral (22.1 vs 34.7%; P < 0.0001).
Comorbidity and late referral
Significantly fewer patients who had been referred late (<90 days) were assessed as having no comorbidity compared with the group referred earlier (39.5 vs 44.5%; P = 0.0046). In terms of specific comorbidities, peripheral vascular disease was significantly less common in the group referred late. On the other hand, liver disease and malignancy were significantly more common in those referred late, perhaps because of the potential for rapid decompensation in these conditions (Table 3.14).
|
Haemoglobin and late referral
Patients referred late had a significantly lower haemoglobin level at dialysis initiation than patients referred earlier (9.4 g/dl vs 10.3; P < 0.001), presumably because of inadequate pre-dialysis care, and the lack of opportunity to optimize anaemia management.
Renal function at the time of starting RRT
Using the abbreviated four variable MDRD calculation, the eGFR of patients starting RRT was calculated. Data from patients with no available creatinine measurement within 14 days before the start of RRT were not used. Patients with an eGFR >20 ml/min/1.73 m2 were excluded from analysis. Data from one centre (Hammersmith and Charing Cross) were excluded from analysis because of errors in the data extraction process of this item. The log of the eGFR was taken to normalize the data, and a two-sample t-test was used to compare the means of the log(eGFR) of those patients with early referral against those with late referral (<3 months).
eGFR and late referral
Estimated GFR was slightly lower in patients referred late compared with earlier referrals (7.34 vs 7.58 ml/min/1.73 m2; P = 0.045). In those over the age of 65 at the time of dialysis initiation the difference was more pronounced (7.41 vs 7.99 ml/min/1.73 m2; P = 0.0003). In Whites only, the difference between late and earlier referrals remained significant but there was little difference in Asians or in Blacks. There were no significant differences in eGFR between those referred late and those referred earlier when stratified by gender, Townsend score or primary renal disease, except that eGFR was significantly lower in patients with renal disease of uncertain aetiology who had been referred late rather than early (6.86 vs 7.40 ml/min/1.73 m2; P = 0.02). When stratifying by comorbidity there were no significant differences in eGFR between the referral groups except that amongst smokers eGFR was significantly lower in those who had been referred late rather than early (7.27 vs 7.95 ml/min/1.73 m2; P = 0.03).
eGFR and age
Older patient groups appear to have a higher geometric mean eGFR at start of dialysis than younger groups (Figure 3.11).
|
Changes over time in eGFR at start of RRT
Analysis of serial data shows a small rise in median eGFR prior to start of RRT in the period 2000–2003 which now appears to have reached a plateau for the last 3 years (Table 3.15).
|
There appears to have been a small increase in eGFR at start of RRT between 1997 and 2003, since when it has remained stable (Figure 3.12). There is no consistent difference between dialysis modalities in eGFR at start of RRT (Figure 3.12).
|
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  R. Rao, D. Ansell, J. A. Gilg, S. J. Davies, E. J. Lamb, and C. R. V. Tomson Effect of change in renal replacement therapy modality on laboratory variables: a cohort study from the UK Renal Registry Nephrol. Dial. Transplant., September 1, 2009; 24(9): 2877 - 2882. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. F. Morrone, S. Mazzaferro, D. Russo, F. Aucella, M. Cozzolino, M. G. Facchini, A. Galfre, F. Malberti, M. C. Mereu, M. Nordio, et al. Interaction between parathyroid hormone and the Charlson comorbidity index on survival of incident haemodialysis patients Nephrol. Dial. Transplant., April 15, 2009; (2009) gfp170v1. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||