Chapter 9 Haemoglobin, ferritin and erythropoietin amongst patients receiving dialysis in the UK in 2007: national and centre-specific analyses

Transcription

1 Haemoglobin, ferritin and erythropoietin amongst patients receiving dialysis in the UK in 2007: national and centre-specific analyses Donald Richardson a, Daniel Ford b, Julie Gilg b and Andrew J Williams c a York District General Hospital, York, UK; b UK Renal Registry, Bristol, UK; c Morriston Hospital, Swansea, UK Key Words Anaemia. Chronic kidney disease. Dialysis. End stage renal disease. Erythropoietin. Ferritin. Haemoglobin. Renal Registry. Epidemiology. Quality improvement Abstract Background: The UK Renal Association (RA) and National Institute for Health and Clinical Excellence (NICE) have published Clinical Practice Guidelines which include recommendations for management of anaemia in established renal failure. Aims: To determine the extent to which the guidelines for anaemia management are met in the UK. Methods: Quarterly data (haemoglobin (Hb) and factors that influence Hb) extracts from renal centres in England, Wales and Northern Ireland (EWNI), and annual data from the Scottish Renal Registry for incident and prevalent renal replacement therapy (RRT) cohorts for 2007 were analysed by the UK Renal Registry (UKRR). Results: In the UK, in % of patients commenced dialysis therapy with Hb g/dl (median Hb 10.3 g/dl). Of incident patients 81% and 87% had a Hb g/dl by 3 and 6 months of dialysis treatment respectively. The median Hb of haemodialysis (HD) patients was 11.6 g/dl with an interquartile range (IQR) of g/dl. Of HD patients 86% had a Hb g/dl. The median Hb of peritoneal dialysis (PD) patients in the UK was 11.9 g/dl (IQR g/dl). 91% of UK PD patients had a Hb g/dl. The median ferritin in HD patients in EWNI was 417 mg/l (IQR 2 598) and 95% of HD patients had a ferritin 5 mg/l. The median ferritin in PD patients was 255 mg/l (IQR ) with 85% of PD patients having a ferritin 5 mg/l. In EWNI the mean ESA dose was higher for HD than PD patients (9,300 vs. 6, IU/week). Conclusions: This year for the first time there has been a small fall (from 85.9% in 2006 to 85.6%) in the percentage of HD patients with an Hb of 510 g/dl. This contrasts with previous annual improvements in this figure and is related to implementation of the new Hb Standard which has a target range of g/dl. Introduction This chapter describes data reported to the UKRR relating to management of renal anaemia during

2 The UK Renal Registry The Eleventh Annual Report The chapter reports outcomes of submitted variables and analyses of these variables in the context of established guidelines and recommendations. The renal national service framework (NSF) part one [1] and the RA minimum standards document 3rd edition [2] state that individuals with chronic kidney disease (CKD) should achieve a Hb of at least 10 g/dl within 6 months of being seen by a nephrologist, unless there is a specific reason why it could not be achieved. The UKRR does not collect a specific measurement from patients 6 months after meeting a nephrologist. Some indication of the standard comes from the Hb of the incident patient population (i.e. the Hb at the start of dialysis). The European Best Practice Guidelines (EBPG) [3] set a minimum target of 11 g/dl but suggest not to go higher than 12 g/dl in severe cardiovascular disease. The United States Kidney Disease Outcomes Quality Initiative (KDOQI) [4] guidelines set a target Hb range of g/dl with a recommendation that the Hb target should not be greater than 13.0 g/dl. The NICE guidelines published in 2006 [5] and the 4th edition of the RA Clinical Practice Guidelines 2007 [6] recommended an outcome Hb of between 10.5 and 12.5 g/dl (with ESA dose changes considered at 11 and 12 g/dl) which allows for the difficulty in consistently narrowing the distribution to between 11 and 12 g/dl. In the 2006 Report much of the data collection had preceded the publication of the NICE guidelines. The 2007 Report begins to show how the attempt to comply with both the g/dl range and the minimum standard of Hb g/dl has impacted on the performance against a combination of measures. The risks associated with low (<10 g/dl) and high (>13 g/dl) Hb are not necessarily equivalent. National and international recommendations for target iron status in CKD remained unchanged from the 2006 Report. The 2007 Renal Association (RA) Clinical Practice Guidelines document, revised European Best Practice Guidelines (EBPGII), Dialysis Outcomes Quality Initiative (DOQI) guidelines and UK NICE anaemia guidelines all recommend a target serum ferritin greater than mg/l and percentage transferrin saturation (TSAT) of more than 20% in patients with CKD. RA guidelines and EBPGII recommend hypochromic red cells (HRC) less than 10%. In addition, EBPGII recommends a target reticulocyte Hb content (CHr) greater than 29 pg/cell. KDOQI recommends a serum ferritin >200 mg/l for HD patients. The NICE guidelines suggest that a hypochromic red cell value >6% suggests ongoing iron deficiency (HRC). To achieve adequate iron status across a patient population, RA guidelines and EBPGII advocate population target medians for ferritin of mg/l, for TSAT of 30 40%, for hypochromic red cells of <2.5% and CHr of 35 pg/cell. EBPGII comments that a serum ferritin target for the treatment population of mg/l ensures that 85 % of patients attain a serum ferritin of mg/l. All guidelines advise that serum ferritin levels should not exceed 0 mg/l since the risk of toxicity increases without conferring additional benefit. The KDOQI and NICE guidelines advise against intravenous iron administration to patients with a ferritin >0 mg/l. Serum ferritin has some disadvantages as an index of iron status. It measures storage iron rather than available iron; behaves as an acute phase reactant and is therefore increased in inflammatory states, malignancy and liver disease; and may not accurately reflect iron stores if measured within a week of the administration of intravenous iron. Of the alternative measures of iron status available, HRC and CHr are generally considered superior to TSAT. Both however require specialised analysers to which few UK renal centres have easy access. Since TSAT is measured infrequently in many centres and most UK centres continue to use serum ferritin for routine iron management, ferritin remains the chosen index of iron status for this report. Methods The incident and prevalent RRT cohorts for 2007 were analysed. The UKRR extracted quarterly data electronically from renal centres in EWNI, data were sent annually from the Scottish Renal Registry. Patients treated with dialysis during the last quarter of 2007 were included in the prevalent analysis if they had been on the same modality of dialysis in the same centre for 3 months. The last available measurement of Hb from each patient from the last two quarters of 2007 was used for analysis. For the incident patient analyses, data from the first quarter after starting dialysis was used. Patients commencing RRT on PD or HD were included. Those receiving a pre-emptive transplant were excluded. Patients were analysed as a complete cohort and divided by modality into groups. The last available ferritin measurement was taken from the last three quarters of the year and analysed for prevalent patients. Ferritin data were only received for three patients from Scotland, so all Scottish centres were excluded from ferritin analyses. 1

3 Anaemia management in UK dialysis patients The completeness of data items were analysed at both centre and country level. All patients were included in analyses but centres with less than % completeness were excluded from the caterpillar and funnel plots showing centre performance. s providing relevant data from less than 20 patients were also excluded from the plots. The number preceding the centre name in each figure indicates the percentage of missing data for that centre. The data were analysed to calculate summary statistics. These were maximum, minimum and average (mean and median) values. Standard deviations and quartile ranges were also found. These data are represented as caterpillar plots showing median values and quartile ranges. The percentage achieving RA and other standards was also calculated for Hb. The percentage of patients achieving serum ferritin 5 mg/l and 5200 mg/l were also calculated. These are represented as caterpillar plots with 95% confidence intervals shown. For the percentage achieving standards, chi-squared values have also been calculated to identify significant variability between centres and between nations. Longitudinal analysis has also been done to calculate overall changes in achievement of standards from 1998 to The UK RA Clinical Practice [1,6] and NICE [5] guidelines in operation at the time these data were collected were as follows: Patients with CKD should achieve a Hb of at least 10 g/dl within 6 months of being seen by a nephrologist, unless there is a specific reason why it could not be achieved. Patients with CKD treated with RRT should have a Hb of between 10.5 and 12.5 g/dl. Patients with CKD should have a serum ferritin greater than g/l and percentage transferrin saturation (TSAT) of more than 20%. Serum ferritin levels in patients with CKD should not exceed 0 g/l. Data regarding ESAs were collected from all centres. s were excluded if fewer than % of patients were on the ESA file. s reporting that fewer than % of HD patients or fewer than 65% of PD patients were treated with ESAs were considered to have incomplete data and were also excluded from further analysis. It is recognised that these exclusion criteria are relatively arbitrary but are in part based upon the frequency distribution graph of centres doses. The UK percentage of patients on ESAs is calculated from this data and incomplete data returns risk seriously impacting on any conclusions drawn. Scotland is excluded from the analysis as there were ESA data returns for only 2 patients. Data are presented as weekly erythropoietin dose. Doses of darbepoietin were harmonised with erythropoietin data by multiplying by 200 and correcting for frequency of administration less than weekly. No adjustments were made with respect to route of administration. The ESA data were collected electronically from renal IT systems but in contrast to laboratory linked variables the ESA dose required manual data entry. The reliability depended upon who entered the data, whether the entry was linked to the prescription or whether the prescriptions were provided by the primary care physician. In the latter case doses may not be as reliably updated as the link between data entry and prescription is indirect. Results Haemoglobin Haemoglobin in incident dialysis patients The Hb at the time of starting RRT gives the only indication of concordance with current anaemia management recommendations in the pre-dialysis (CKD 5 not yet on dialysis) group. Patients for conservative care of end stage renal failure were by definition excluded from the dataset. The UKRR plans to collect and report CKD 5 data from patients who subsequently commence RRT as well as those managed conservatively. The percentage of data returned and outcome Hb are listed in table 9.1. The median Hb of patients at the time of starting dialysis in the UK was 10.3 g/dl with 58% of patients having a Hb g/dl (vs. % for 2006 Report). The variation between centres remained high (35 88%). The median starting Hb is shown in figure 9.1 and the percentage starting with a Hb g/dl by centre is given in figure 9.2. The distribution of Hb in incident dialysis patients during 2007 is shown in figure 9.3. The median Hb and the percentage of incident dialysis patients in 2006 with Hb g/dl by time on dialysis are shown in figures 9.4 and 9.5. The annual distribution (figure 9.6) of Hb in incident dialysis patients has not changed significantly since Haemoglobin in prevalent haemodialysis patients The compliance with data returns and Hb outcome for prevalent HD patients are shown in table 9.2. The median Hb of patients on HD in the UK was 11.6 g/dl with an IQR of g/dl. In the UK, 86% of HD patients had a Hb g/dl. The median Hb by centre, compliance with the previous UK minimum standard of Hb g/dl and EBPG standard of Hb g/dl are shown in figures 9.7, 9.8 and 9.9 respectively. The distribution of Hb in HD patients by centre is shown in figure The compliance with the NICE and RA Clinical Practice Guidelines recommended range of g/dl is shown in figure The majority of centres complied well with respect to both outcomes but it was possible to fall within 2 3 sd limits of the mean in the funnel 151

4 The UK Renal Registry The Eleventh Annual Report Table 9.1. Haemoglobin data for new patients starting haemodialysis or peritoneal dialysis during 2007 % data return Median Hb g/dl % range Inter-quartile range % Hb 510 g/dl Abrdn Airdrie 18 Antrim B Heart B QEH Bangor Basldn Belfast Bradfd Brightn Bristol Camb Cardff Carlis Carsh Chelms Clwyd 87 Covnt D & Gall 94 Derby Derry Donc Dorset Dudley Dundee Dunfn 9 Edinb 0 Exeter Glasgw Glouc Hull Inverns 81 Ipswi Klmarnk L Barts L Guys L Kings L Rfree L West Leeds Leic Liv Ain Liv RI M Hope M RI Middlbr Newc Newry Norwch Nottm Oxford Plymth Ports Prestn Redng Sheff Shrew Stevng

5 Anaemia management in UK dialysis patients Table 9.1. Continued % data return Median Hb g/dl % range Inter-quartile range % Hb 510 g/dl Sthend Stoke Sund Swanse Truro Tyrone Ulster 95 Wirral Wolve Wrexm York England N Ireland Scotland Wales UK Blank cells denote centres excluded from analyses due to low patient numbers or poor data completeness N = 5,084 Upper quartile Median Hb Lower quartile Haemoglobin g/dl L West 38 Klmarnk 6 Chelms 3 Wolve 25 L Rfree 3 York 0 Oxford 3 Newc 1 Cardff 21 Plymth 21 Antrim 9 Derby 0 Truro 25 Wirral 5 Covnt 24 Camb 5 Liv RI 4 Wrexm 4 L Kings 24 M RI 13 Dorset 0 Swanse 0 Nottm 0 Bangor 0 Carlis 30 Dundee 0 Ports 0 Sund 1 Redng 2 Carsh 0 Stevng 1 Middlbr 2 Abrdn 3 Brightn 1 Bristol 14 Belfast 5 Bradfd 0 Basldn 0 Glouc 10 B Heart 3 M Hope 22 L Guys 0 Leeds 7 Norwch 0 Leic 9 Ipswi 0 Sthend 5 Hull 0 Sheff 33 B QEH 5 Prestn 3 Dudley 10 Glasgw 0 L Barts 5 Tyrone 1 Exeter 3 Liv Ain 9 England 13 N Ireland 40 Scotland 1 Wales 12 UK Fig Median haemoglobin for incident dialysis patients at start of dialysis treatment in N = 5,084 Upper 95% Cl % with Hb >10 g/dl Lower 95% Cl 33 L West 38 Klmarnk 0 Oxford 25 L Rfree 21 Antrim 13 Dorset 6 Chelms 3 Newc 25 Wirral 0 Truro 1 Cardff 3 Wolve 21 Plymth 9 Derby 3 York 0 Carlis 5 Liv RI 5 Covnt 24 M RI 2 Carsh 0 Nottm 4 L Kings 1 Redng 0 Sund 14 Belfast 30 Dundee 0 Stevng 0 Swanse 24 Camb 4 Wrexm 0 Bangor 0 Basldn 0 Ports 0 Glouc 1 Middlbr 0 Leic 2 Abrdn 1 Bristol 22 L Guys 3 Brightn 0 Leeds 5 Bradfd 7 Norwch 10 B Heart 3 M Hope 9 Ipswi 0 Sheff 5 Prestn 0 Sthend 5 Hull 33 B QEH 0 L Barts 3 Dudley 10 Glasgw 5 Tyrone 1 Exeter 3 Liv Ain 9 England 13 N Ireland 40 Scotland 1 Wales 12 UK Fig Percentage of incident dialysis patients with Hb 510 g/dl at start of dialysis treatment in

6 The UK Renal Registry The Eleventh Annual Report 12 UK 1 Wales 40 Scotland 13 N Ireland 9 England 3 Liv Ain 1 Exeter 5 Tyrone 10 Glasgw 3 Dudley 0 L Barts 33 B QEH 5 Hull 0 Sthend 5 Prestn 0 Sheff 9 Ipswi 3 M Hope 10 B Heart 7 Norwch 5 Bradfd 0 Leeds 3 Brightn 22 L Guys 1 Bristol 2 Abrdn 0 Leic 1 Middlbr 0 Glouc 0 Ports 0 Basldn 0 Bangor 4 Wrexm 24 Camb 0 Swanse 0 Stevng 30 Dundee 14 Belfast 0 Sund 1 Redng 4 L Kings 0 Nottm 2 Carsh 24 M RI 5 Covnt 5 Liv RI 0 Carlis 3 York 9 Derby 21 Plymth 3 Wolve 1 Cardff 25 Wirral 0 Truro 3 Newc 6 Chelms 13 Dorset 21 Antrim 25 L Rfree 0 Oxford 38 Klmarnk 33 L West <9.0 (Hb, g/dl) > Fig Distribution of haemoglobin in incident dialysis patients at start of dialysis treatment in

7 Anaemia management in UK dialysis patients Haemoglobin g/dl Start 3 months 6 months 12 months Time since commencing dialysis Upper quartile Median Hb Lower quartile Fig Median haemoglobin, by time on dialysis, for incident dialysis patients in Start 3 months 6 months 12 months Time since commencing dialysis Upper 95% Cl % with Hb >10 g/dl Lower 95% Cl Fig Percentage of incident dialysis patients in 2006 with Hb 510 g/dl, by time on dialysis Percentage of incident patients >12 (Hb, g/dl) < Year of start Fig Distribution of haemoglobin in incident dialysis patients by year of start 14 N = 15,858 Upper quartile Median Hb Lower quartile Haemoglobin g/dl Liv RI 15 Inverns 0 Wolve 3 Leeds 23 L Rfree 0 Derby 1 Wrexm 2 D&Gall 5 Carlis 2 Chelms 3 Liv Ain 0 Stevng 2 Middlbr 4 B QEH 17 Carsh 2 Nottm 2 Hull 10 B Heart 27 M RI 10 Dudley 1 Swanse 5 Klmarnk 3 Airdrie 2 Abrdn 3 Dundee 14 M Hope 2 Antrim 2 Newry 3 Tyrone 0 Dorset 1 Truro 0 Glouc 2 Sund 3 Prestn 2 Cardff 1 Leic 10 Belfast 0 Derry 2 Bradfd 24 Wirral 2 Basldn 0 Ipswi 1 Sheff 1 Oxford 0 Donc 3 Bangor 3 Plymth 7 Clwyd 1 Ulster 42 Camb 0 Brightn 1 Ports 1 Covnt 9 Norwch 2 Glasgw 3 L Guys 0 L Barts 3 Sthend 9 England 5 N Ireland 22 Scotland 2 Wales 10 UK Fig Median haemoglobin in patients treated with HD 155

8 The UK Renal Registry The Eleventh Annual Report Table 9.2. Haemoglobin data for prevalent HD patients % data return Median Hb g/dl % range Inter-quartile range Mean Hb g/dl Standard deviation % with Hb 510 g/dl % with Hb 511 g/dl Abrdn Airdrie Antrim B Heart B QEH Bangor Basldn Belfast Bradfd Brightn Bristol Camb Cardff Carlis Carsh Chelms Clwyd Covnt D & Gall Derby Derry Donc Dorset Dudley Dundee Dunfn 25 Edinb 1 Exeter Glasgw Glouc Hull Inverns Ipswi Klmarnk L Barts L Guys L Kings L Rfree L West 44 Leeds Leic Liv Ain Liv RI M Hope M RI Middlbr Newc Newry Norwch

9 Anaemia management in UK dialysis patients Table 9.2. Continued % data return Median Hb g/dl % range Inter-quartile range Mean Hb g/dl Standard deviation % with Hb 510 g/dl % with Hb 511 g/dl Nottm Oxford Plymth Ports Prestn Redng Sheff Shrew Stevng Sthend Stoke Sund Swanse Truro Tyrone Ulster Wirral Wolve Wrexm York England N Ireland Scotland Wales UK Blank cells denote centres excluded from analyses due to low patient numbers or poor data completeness N = 15,858 Upper 95% Cl % with Hb >10 g/dl Lower 95% Cl 65 5 Carlis 15 Inverns 0 Derby 3 Liv Ain 2 D&Gall 24 Wirral 2 Hull 2 Middlbr 23 L Rfree 6 Liv RI 3 Tyrone 1 Truro 9 Norwch 2 Antrim 1 Wrexm 0 Stevng 1 Swanse 3 Leeds 3 Bangor 3 Plymth 0 Ipswi 2 Newry 3 Prestn 2 Nottm 2 Abrdn 0 Glouc 10 B Heart 2 Cardff 0 Wolve 3 Airdrie 4 B QEH 2 Chelms 17 Carsh 1 Oxford 1 Sheff 0 Derry 5 Klmarnk 2 Sund 14 M Hope 2 Bradfd 1 Covnt 1 Leic 0 Dorset 3 Dundee 7 Clwyd 2 Glasgw 0 Brightn 27 M RI 0 Donc 2 Basldn 1 Ports 3 L Guys 1 Ulster 10 Belfast 42 Camb 10 Dudley 3 Sthend 0 L Barts 9 England 5 N Ireland 22 Scotland 2 Wales 10 UK Fig Percentage of HD patients with Hb 510 g/dl 157

10 The UK Renal Registry The Eleventh Annual Report N = 15,858 Upper 95% Cl % with Hb >11 g/dl Lower 95% Cl Liv RI 0 Derby 15 Inverns 2 D&Gall 2 Chelms 3 Leeds 5 Carlis 23 L Rfree 2 Antrim 2 Middlbr 3 Liv Ain 0 Wolve 1 Swanse 2 Hull 2 Newry 3 Tyrone 0 Stevng 10 B Heart 3 Prestn 1 Truro 2 Abrdn 2 Nottm 0 Ipswi 24 Wirral 17 Carsh 1 Wrexm 2 Sund 4 B QEH 2 Bradfd 0 Glouc 1 Oxford 3 Airdrie 3 Dundee 14 M Hope 0 Dorset 3 Plymth 27 M RI 1 Sheff 1 Leic 2 Cardff 5 Klmarnk 3 Bangor 10 Dudley 0 Derry 10 Belfast 9 Norwch 0 Brightn 0 Donc 1 Covnt 2 Glasgw 2 Basldn 42 Camb 3 L Guys 7 Clwyd 1 Ports 0 L Barts 1 Ulster 3 Sthend 9 England 5 N Ireland 22 Scotland 2 Wales 10 UK Fig Percentage of HD patients with Hb 511 g/dl plot (figure 9.12) for percentage of patients with Hb and g/dl and yet have a poor compliance with percentage of Hb g/dl (figure 9.13). This demonstrated that compliance with one standard (Hb and g/dl) can be achieved without compliance with another standard (Hb g/dl). Figures 9.12 and 9.13 should be used in conjunction with table 9.3 to identify centres. Haemoglobin in prevalent peritoneal dialysis patients In the UK 91% of patients on PD had a Hb g/dl (table 9.4). The median Hb of patients on PD in the UK was 11.9 g/dl with an IQR of g/dl (table 9.4). The median Hb by centre, compliance with the UK minimum standard Hb g/dl and EBPG Hb g/dl are shown in figures 9.14, 9.15 and 9.16 respectively. The compliance with recommended range Hb and g/dl (NICE & RA) is shown in figure The distribution of Hb in PD patients by centre is shown in figure The funnel plot for percentage Hb g/dl is shown in figure 9.19 which can be used in conjunction with table 9.5 to identify centres. Relationship between Hb in incident and prevalent dialysis patients in 2007 The relationship between the percentage of new and prevalent dialysis (HD and PD) patients with a Hb g/dl is demonstrated in figure Correlation between median haemoglobin and compliance with clinical guidelines The use of Rose-Day plots demonstrated the relationship between the population mean (and standard deviation) and the compliance with minimum standards. The plots for Hb g/dl and g/dl for HD and PD populations are given in figures 9.21 to The compliance with minimum standards over time between 1998 and 2007 are shown in figure 9.25 for prevalent patients and in figure 9.26 for incident and prevalent patients between 1998 and Changes in haemoglobin by length of time on renal replacement therapy over time The median Hb of patients treated with HD increased during the first year of treatment (figure 9.27) but did not do so in patients treated with PD (figure 9.28). The median Hb of HD patients in 2007 (figure 9.27) was lower than in 2006 irrespective of time on RRT. The Hb in PD patients (figure 9.28) had been stable for some years and remained higher than in HD patients. 158

11 Anaemia management in UK dialysis patients 10 UK 2 Wales 22 Scotland 5 N Ireland 9 England 0 L Barts 3 Sthend 10 Dudley 42 Camb 10 Belfast 1 Ulster 3 L Guys 1 Ports 2 Basldn 0 Donc 27 M RI 0 Brightn 2 Glasgw 7 Clwyd 3 Dundee 0 Dorset 1 Leic 1 Covnt 2 Bradfd 14 M Hope 2 Sund 5 Klmarnk 0 Derry 1 Sheff 1 Oxford 17 Carsh 2 Chelms 4 B QEH 3 Airdrie 0 Wolve 2 Cardff 10 B Heart 0 Glouc 2 Abrdn 2 Nottm 3 Prestn 0 Ipswi 3 Plymth 3 Bangor 3 Leeds 1 Swanse 2 Newry 0 Stevng 1 Wrexm 2 Antrim 9 Norwch 1 Truro 3 Tyrone 6 Liv RI 23 L Rfree 2 Middlbr 2 Hull 24 Wirral 2 D&Gall 3 Liv Ain 0 Derby 15 Inverns 5 Carlis <9.0 (Hb, g/dl) > Fig Distribution of haemoglobin in patients treated with HD 159

12 The UK Renal Registry The Eleventh Annual Report N = 15,858 Upper 95% Cl % with Hb g/dl Lower 95% Cl Newry 0 Ipswi 5 Carlis 2 Antrim 1 Truro 3 Plymth 3 Tyrone 0 Stevng 2 Bradfd 1 Ulster 9 Norwch 0 Derry 24 Wirral 2 Hull 1 Swanse 2 Chelms 1 Sheff 0 Brightn 2 Nottm 3 Prestn 3 Bangor 7 Clwyd 10 B Heart 3 L Guys 2 Glasgw 2 D&Gall 2 Cardff 17 Carsh 3 Airdrie 0 Glouc 1 Leic 42 Camb 2 Abrdn 2 Middlbr 15 Inverns 14 M Hope 3 Dundee 5 Klmarnk 1 Ports 0 Dorset 23 L Rfree 10 Belfast 3 Sthend 1 Covnt 1 Oxford 0 Derby 4 B QEH 0 Donc 3 Liv Ain 3 Leeds 2 Sund 0 L Barts 6 Liv RI 1 Wrexm 2 Basldn 27 M RI 10 Dudley 0 Wolve 9 England 5 N Ireland 22 Scotland 2 Wales 10 UK Fig Percentage of HD patients with Hb and g/dl 65 Solid lines show 95% confidence limits Dotted lines show 99.9% confidence limits 95 Solid lines show 95% confidence limits Dotted lines show 99.9% confidence limits Number of patients with data in centre Fig Funnel plot of percentage of HD patients with Hb and g/dl Number of patients with data in centre Fig Funnel plot of percentage of HD patients with Hb 510 g/dl 1

13 Anaemia management in UK dialysis patients Table 9.3. Percentage of HD patients achieving Hb 510 g/dl and Hb g/dl N with Hb % with Hb 510 g/dl % Hb g/dl N with Hb % with Hb 510 g/dl % Hb g/dl Derry Newc D & Gall Redng Donc Norwch Bangor Belfast Clwyd M RI Tyrone Stoke Inverns Wolve Ulster Middlbr Wrexm Exeter Carlis M Hope Newry Swanse Ipswi Covnt Chelms Hull Dudley Brightn York Stevng Liv Ain L Kings Sthend B Heart Plymth Oxford Klmarnk Nottm Antrim Liv RI Basldn Ports 371 Wirral Prestn Airdrie Bristol Dorset L Guys Truro Carsh Sund L Rfree Dundee Cardff Shrew Leeds Bradfd Sheff Glouc Glasgw Derby L Barts Camb Leic Abrdn B QEH Table 9.4. Haemoglobin data for prevalent PD patients % data return Median Hb g/dl % range Inter-quartile range Mean Hb g/dl Standard deviation % with Hb 510 g/dl % with Hb 511 g/dl Abrdn Airdrie 82 Antrim 94 B Heart B QEH Bangor Basldn Belfast Bradfd Brightn Bristol Camb Cardff Carlis Carsh Chelms Clwyd 92 Covnt D & Gall 161

14 The UK Renal Registry The Eleventh Annual Report Table 9.4. Continued % data return Median Hb g/dl % range Inter-quartile range Mean Hb g/dl Standard deviation % with Hb 510 g/dl % with Hb 511 g/dl Derby Derry Donc Dorset Dudley Dundee Dunfn 17 Edinb 0 Exeter Glasgw Glouc Hull Inverns 0 Ipswi Klmarnk L Barts L Guys L Kings L Rfree L West 5 Leeds Leic Liv Ain n/a Liv RI M Hope M RI Middlbr Newc Newry Norwch Nottm Oxford Plymth Ports Prestn Redng Sheff Shrew Stevng Sthend 94 Stoke Sund Swanse Truro Tyrone Ulster Wirral Wolve Wrexm York England N Ireland Scotland Wales UK Blank cells denote centres excluded from analyses due to low patient numbers or poor data completeness 162

15 Anaemia management in UK dialysis patients Haemoglobin g/dl N = 3,758 Upper quartile Median Hb Lower quartile 10 3 Chelms 2 Wolve 0 Stevng 8 Middlbr 3 B Heart 0 Basldn 4 Norwch 6 Dudley 6 Hull 16 Plymth 1 Cardff 2 Dorset 1 Ports 1 L Barts 1 Leeds 8 Liv RI 0 Bangor 6 Covnt 10 Wrexm 3 Abrdn 4 Dundee 1 Leic 2 Ipswi 0 Camb 1 M RI 0 Bradfd 0 Glouc 5 M Hope 5 Klmarnk 0 Nottm 0 Derby 0 Brightn 2 L Guys 0 Donc 4 Swanse 29 Wirral 0 Truro 3 Prestn 0 Oxford 3 Glasgw 11 L Rfree 0 Sheff 14 B QEH 4 Belfast 4 Carsh 5 England 3 N Ireland 38 Scotland 3 Wales 7 UK Fig Median haemoglobin in patients treated with PD Upper 95% Cl % with Hb >10 g/dl N = 3,758 Lower 95% Cl 0 Basldn 0 Bangor 0 Glouc 3 Abrdn 4 Belfast 10 Wrexm 11 L Rfree 0 Brightn 8 Middlbr 0 Truro 2 Ipswi 29 Wirral 2 Dorset 3 Chelms 16 Plymth 0 Camb 3 B Heart 4 Swanse 4 Norwch 5 Klmarnk 2 Wolve 14 B QEH 0 Stevng 0 Nottm 8 Liv RI 0 Derby 6 Hull 5 M Hope 0 Donc 1 Cardff 6 Covnt 1 Ports 1 Leic 0 Bradfd 4 Dundee 1 M RI 3 Prestn 0 Sheff 4 Carsh 0 Oxford 6 Dudley 1 Leeds 1 L Barts 3 Glasgw 2 L Guys 5 England 3 N Ireland 38 Scotland 3 Wales 7 UK Fig Percentage of PD patients with Hb 510 g/dl 40 Upper 95% Cl % with Hb >11 g/dl Lower 95% Cl N = 3,758 0 Basldn 3 Abrdn 10 Wrexm 3 Chelms 8 Middlbr 0 Bangor 3 B Heart 0 Stevng 4 Norwch 0 Derby 2 Wolve 0 Brightn 8 Liv RI 6 Hull 2 Dorset 1 Leeds 0 Camb 0 Truro 16 Plymth 4 Dundee 5 Klmarnk 1 Cardff 0 Glouc 1 Leic 4 Belfast 0 Nottm 4 Swanse 1 Ports 0 Donc 11 L Rfree 6 Covnt 29 Wirral 3 Prestn 1 M RI 3 Glasgw 2 Ipswi 14 B QEH 0 Oxford 1 L Barts 2 L Guys 0 Sheff 4 Carsh 0 Bradfd 6 Dudley 5 M Hope 5 England 3 N Ireland 38 Scotland 3 Wales 7 UK Fig Percentage of PD patients with Hb 511 g/dl 163

16 The UK Renal Registry The Eleventh Annual Report N = 3,758 Upper 95% Cl % with Hb g/dl Lower 95% Cl 24 Wirral 1 Wrexm 0 Glouc 23 L Rfree 2 Abrdn 1 Truro 5 Klmarnk 42 Camb 1 Sheff 3 Bangor 6 Liv RI 2 Bradfd 2 Basldn 0 Ipswi 10 B Heart 2 Middlbr 10 Belfast 0 Derby 1 Swanse 2 Nottm 3 Prestn 17 Carsh 3 L Guys 3 Dundee 0 Brightn 0 Dorset 1 Covnt 4 B QEH 3 Plymth 2 Glasgw 0 Donc 1 Leic 27 M RI 9 Norwch 1 Ports 0 Stevng 3 Leeds 1 Oxford 2 Cardff 14 M Hope 2 Hull 0 L Barts 0 Wolve 10 Dudley 2 Chelms 9 England 5 N Ireland 22 Scotland 2 Wales 10 UK Fig Percentage of PD patients with Hb and g/dl 164

17 Anaemia management in UK dialysis patients 7 UK 3 Wales 38 Scotland 3 N Ireland 5 England 2 L Guys 3 Glasgw 1 L Barts 1 Leeds 6 Dudley 0 Oxford 4 Carsh 0 Sheff 3 Prestn 1 M RI 0 Bradfd 4 Dundee 1 Leic 1 Ports 6 Covnt 1 Cardff 5 M Hope 0 Donc 6 Hull 0 Derby 8 Liv RI 0 Nottm 0 Stevng 14 B QEH 2 Wolve 5 Klmarnk 4 Norwch 4 Swanse 3 B Heart 0 Camb 16 Plymth 3 Chelms 2 Dorset 29 Wirral 2 Ipswi 0 Truro 8 Middlbr 0 Brightn 11 L Rfree 10 Wrexm 4 Belfast 3 Abrdn 0 Glouc 0 Bangor 0 Basldn <9.0 (Hb, g/dl) > Fig Distribution of haemoglobin in patients treated with PD 165

18 The UK Renal Registry The Eleventh Annual Report Solid lines show 95% confidence limits Dotted lines show 99.9% confidence limits Number of patients with data in centre Fig Funnel plot of percentage of PD patients with Hb 510 g/dl Table 9.5. Percentage of PD patients achieving Hb 510 g/dl N with Hb % with Hb 510 g/dl N with Hb % with Hb 510 g/dl Wirral Covnt 61 Middlbr Exeter 91 Truro Derby York 23 Swanse Basldn 26 Bristol Dundee L Kings Wrexm Prestn Abrdn Brightn B Heart Hull Glouc Liv RI Bangor Redng 86 Plymth Sheff Donc Glasgw Shrew Stoke 89 Chelms Ports 91 Bradfd Leeds Stevng B QEH Klmarnk L Rfree Ipswi Carsh Newc M Hope Camb M RI Dudley Oxford Wolve Nottm Belfast Cardff 146 Dorset Leic Norwch L Barts L Guys

19 Anaemia management in UK dialysis patients New dialysis patients Prevalent dialysis patients Liv Ain Exeter Tyrone Glasgw Dudley L Barts B QEH Hull Sthend Prestn Sheff Ipswi M Hope B Heart Norwch Bradfd Leeds Brightn Bristol L Guys Abrdn Leic Middlbr Glouc Ports Stoke Basldn Wrexm Bangor Camb Swanse Stevng Dundee Belfast Sund Redng L Kings Nottm Carsh M RI Covnt Liv RI York Carlis Derby Plymth Wolve Cardff Wirral Truro Newc Chelms Dorset Antrim L Rfree Oxford Shrew Klmarnk England N Ireland Scotland Wales UK Fig Percentage of new and prevalent dialysis patients with Hb 510 g/dl Percentage Hb >10 g/dl Median Hb g/dl Fig Percentage of HD patients with Hb 510 g/dl plotted against median haemoglobin Percentage Hb >10 g/dl Median Hb g/dl Fig Percentage of PD patients with Hb510 g/dl plotted against median haemoglobin Percentage Hb >11 g/dl Percentage Hb >11 g/dl Median Hb g/dl Fig Percentage of HD patients with Hb 511 g/dl plotted against median haemoglobin Median Hb g/dl Fig Percentage of PD patients with Hb 511 g/dl plotted against median haemoglobin 167

20 The UK Renal Registry The Eleventh Annual Report Upper 95% Cl % with Hb >10 g/dl Lower 95% Cl Haemodialysis Year Peritoneal dialysis Figure Percentage of prevalent HD and PD patients ( ) with Hb 510 g/dl Upper 95% Cl % with Hb >10 g/dl Lower 95% Cl 40 Incident patients Prevalent patients Year Figure Percentage of incident and prevalent dialysis patients ( ) with Hb 510 g/dl Hb g/dl <6 months 6m 1 year 1 2 years Time on RRT >2 years All HD 1999 HD 2000 HD 2001 HD 2002 HD 2003 HD 2004 HD 2005 HD 2006 HD 2007 Figure Median haemoglobin plotted by length of time on RRT (HD patients) 168

21 Anaemia management in UK dialysis patients Hb g/dl <6 months 6m 1 year 1 2 years Time on RRT >2 years All HD 1999 HD 2000 HD 2001 HD 2002 HD 2003 HD 2004 HD 2005 HD 2006 HD 2007 Figure Median haemoglobin plotted by length of time on RRT (PD patients) Factors affecting haemoglobin Ferritin Completeness of ferritin returns for patients treated with HD and PD The completeness of serum ferritin returns to the UKRR is shown in table 9.6. Not all centres used serum ferritin as the sole indicator of iron status. Completeness of data for serum ferritin returned for England, Wales and Northern Ireland improved by comparison with the previous year. For Scotland a lack of an automated biochemistry link to the renal IT system is thought to account for the low rate of return. In other cases of missing data, renal centres may need to address organisational processes in dealing with automatic download facilities to ensure that serum ferritin is checked, or alternatively that a declaration is made that alternative measures of iron status are being utilised. Ferritin in prevalent dialysis patients Percentage returns, serum ferritin concentrations and IQR are presented in tables 9.7 and 9.8 for HD and PD patients respectively. The percentage of patients with a ferritin mg/l by centre for HD and PD patients is shown in table 9.9. The median and IQR for serum ferritin for HD and PD patients by centre are given in figures 9.29 and 9.30 respectively. The percentage of patients with a serum ferritin 5 mg/l, 5200 mg/l and mg/l are shown in figures 9.31, 9.32 and 9.33 for HD and figures 9.34, 9.35 and 9.36 for PD respectively. All centres achieved greater than 75% compliance with a serum of ferritin 5 mg/l for HD patients and all but 5 centres achieved >% compliance. The PD population had a lower median ferritin value (255 mg/l, IQR vs. 417 mg/l, IQR for HD). All but 5 centres achieved greater than % compliance for serum ferritin 5 mg/l in the PD population. Changes in ferritin The compliance with guidelines for ferritin in the HD and PD populations has remained stable over the last 5 years at approximately 95% and 85% respectively. The serial values are shown in figure The difference between the compliance in HD and PD was probably because of the lower requirement for ESA to achieve target Hb levels in the PD population. There was therefore a lower requirement for intravenous iron supplementation. The median serum ferritin outcome over time is shown in figure Ferritin and length of time on renal replacement therapy The median serum ferritin increased during the first year in patients treated with HD and during the first 2 years in those treated with PD (figures 9.39 and 9.40). After 2 years the levels remained stable in both groups of patients. 169

22 The UK Renal Registry The Eleventh Annual Report Table 9.6. Completeness of ferritin returns HD % PD % HD % PD % Abrdn 0 0 L Rfree Airdrie 0 0 L West Antrim 98 Leeds B Heart Leic B QEH Liv Ain 96 n/a Bangor 95 Liv RI Basldn 98 M Hope 93 Belfast M RI 73 Bradfd 99 Middlbr Brightn Newc Bristol Newry 96 Camb 72 Norwch Cardff Nottm 99 Carlis 95 Oxford Carsh Plymth 97 Chelms 97 Ports 95 Clwyd Prestn 99 Covnt Redng D & Gall 0 0 Sheff 99 Derby 99 Shrew Derry Stevng Donc Sthend Dorset 96 Stoke 99 Dudley 87 Sund 97 Dundee 1 0 Swanse Dunfn 0 0 Truro 99 Edinb 1 0 Tyrone 96 Exeter Ulster Glasgw 0 0 Wirral Glouc 98 Wolve Hull Wrexm Inverns 0 Ipswi 82 England Klmarnk 0 0 N Ireland L Barts 99 Scotland 0 0 L Guys Wales L Kings UK

23 Anaemia management in UK dialysis patients Table 9.7. Ferritin in HD patients % data return Median ferritin % range Inter-quartile range % ferritin5 mg/l Antrim B Heart B QEH Bangor Basldn Belfast Bradfd Brightn Bristol Camb Cardff Carlis Carsh Chelms Clwyd Covnt Derby Derry Donc Dorset Dudley Exeter Glouc Hull Ipswi L Barts L Guys L Kings L Rfree L West Leeds Leic Liv Ain Liv RI M Hope M RI Middlbr Newc Newry Norwch Nottm Oxford Plymth Ports Prestn Redng Sheff Shrew Stevng Sthend Stoke

24 The UK Renal Registry The Eleventh Annual Report Table 9.7. Continued % data return Median ferritin % range Inter-quartile range % ferritin5 mg/l Sund Swanse Truro Tyrone Ulster Wirral Wolve Wrexm York England N Ireland Wales E, W & NI Table 9.8. Ferritin in PD patients % data return Median ferritin % range Inter-quartile range % ferritin 5 mg/l Antrim B Heart B QEH Bangor Basldn Belfast Bradfd Brightn Bristol Camb Cardff Carlis Carsh Chelms Clwyd 92 Covnt Derby Derry Donc Dorset Dudley Exeter Glouc Hull Ipswi L Barts L Guys L Kings

25 Anaemia management in UK dialysis patients Table 9.8. Continued % data return Median ferritin % range Inter-quartile range % ferritin 5 mg/l L Rfree L West Leeds Leic Liv Ain n/a Liv RI M Hope M RI Middlbr Newc Newry Norwch Nottm Oxford Plymth Ports Prestn Redng Sheff Shrew Stevng Sthend 94 Stoke Sund Swanse Truro Tyrone Ulster Wirral Wolve Wrexm 13 York England N Ireland Wales E, W & NI Blank cells denote centres excluded from analyses due to low patient numbers or poor data completeness n/a not applicable 173

26 The UK Renal Registry The Eleventh Annual Report Table 9.9. with ferritin mg/l HD PD % ferritin mg/l 95% CI % ferritin mg/l 95% CI Antrim B Heart B QEH Bangor Basldn Belfast Bradfd Brightn Bristol Camb Cardff Carlis Carsh Chelms Clwyd Covnt Derby Derry Donc Dorset Dudley Exeter Glouc Hull Ipswi L Barts L Guys L Kings L Rfree L West Leeds Leic Liv Ain n/a n/a Liv RI M Hope M RI Middlbr Newc Newry Norwch Nottm Oxford Plymth Ports Prestn Redng Sheff Shrew Stevng Sthend Stoke Sund Swanse Truro

27 Anaemia management in UK dialysis patients Table 9.9. Continued HD PD % ferritin mg/l 95% CI % ferritin mg/l 95% CI Tyrone Ulster Wirral Wolve Wrexm York England N Ireland Wales E, W & NI Blank cells denote centres excluded from analyses due to low patient numbers or poor data completeness n/a not applicable 1,400 1,200 1,000 N = 14,793 Upper quartile Median ferritin Lower quartile Ferritin µg/l Tyrone 0 Chelms 0 Prestn 4 Newry 0 Derry 37 Wirral 20 L West 0 Ulster 13 Norwch 4 Belfast 1 Nottm 9 Liv RI 1 Truro 1 Wolve 40 M Hope 2 Leeds 1 Bradfd 2 Glouc 2 Brightn 4 Liv Ain 21 L Rfree 2 Antrim 3 Sund 5 Carlis 0 Dorset 3 Plymth 1 Sheff 4 Cardff 5 Bangor 0 L Barts 2 Donc 2 Hull 1 Stevng 7 Clwyd 27 M RI 4 Middlbr 3 B QEH 3 L Guys 1 Swanse 11 Leic 0 Ipswi 2 Basldn 2 Sthend 0 Dudley 28 Camb 2 Oxford 0 Derby 19 Carsh 2 Covnt 14 Wrexm 7 B Heart 5 Ports 8 England 3 N Ireland 4 Wales 16 E, W & NI Fig Median ferritin in patients treated with HD 0 0 N = 3,516 Upper quartile Median ferritin Lower quartile Ferritin µg/l Wirral 1 Stoke 1 Brightn 0 Bangor 7 Hull 7 Norwch 4 Dorset 0 Nottm 5 Leic 4 L Rfree 1 L Barts 0 Truro 3 Chelms 4 Belfast 0 Bradfd 0 Glouc 1 Derby 8 Liv RI 18 Ipswi 30 L West 0 Sheff 1 Prestn 7 M Hope 0 Camb 3 Oxford 4 Swanse 8 Middlbr 1 Leeds 24 Donc 10 Ports 0 Basldn 13 B QEH 11 Covnt 0 Plymth 5 Stevng 4 Wolve 3 B Heart 13 Dudley 4 Carsh 2 L Guys 3 Cardff 0 M RI 4 England 2 N Ireland 12 Wales 13 E, W & NI Fig Median ferritin in patients treated with PD 175