Reclassification of Chronic Kidney Disease Stage, Eligibility for Cystatin-C and its Associated Costs in a UK Primary Care Cohort

Transcription

1 Reclassification of Chronic Kidney Disease Stage, Eligibility for Cystatin-C and its Associated Costs in a UK Primary Care Cohort Rupert W Major 1,2, David Shepherd 2, Nigel J Brunskill 1,3 1 Department of Nephrology, Leicester General Hospital, Leicester UK 2 Department of Health Sciences, University of Leicester, Leicester, UK 3 Department of Infection, Immunity and Inflammation, University of Leicester, Leicester, UK Word Count: 2,826 Number of References: 23 Abstract: 250 words Short title: Reclassification of CKD and Eligibility for Cystatin-C Correspondence to: Dr Rupert Major Kidney Research UK Academic Clinical Fellow Department of Health Sciences Centre for Medicine 15 Lancaster Road Leicester, LE1 7HA, UK rwlm2@le.ac.uk Tel: Fax:

2 Abstract Background and Aims: CKD-EPI egfr, compared to MDRD, has superior performance in predicting renal, cardiovascular and mortality events. Cystatin-C further improves prediction. Our primary aim was to assess the change in prevalence and classification of CKD in converting from MDRD to CKD-EPI in an unselected primary care CKD population. Our secondary aims were to determine eligibility for cystatin-c testing based on NICE guidance and the associated costs. Methods: egfr data from an unselected UK primary care CKD cohort was studied to assess reclassification of CKD stages from MDRD to CKD-EPI, suitability for cystatin-c testing and its associated cost. Results: 24,660 individuals had 2 MDRD egfr results <60 ml/min/1.73m 2 >3 months apart (7.0% of adult population). Mean age was 75.2 years (SD 11.4 years) with 15,265 (61.9%) females. Mean egfr was 2.88 ml/min/1.73m 2 lower with CKD- EPI egfr versus MDRD egfr (49.7 versus 46.8, t-test p<0.0001, 95% CI 2.85 to 2.91). 12.0% and 1.3% of individuals were re-categorised to a more or less advanced CKD stage respectively. 2.8% of the population were categorised as 3a- A1 CKD-EPI and therefore potentially suitable for cystatin-c. The estimated initial cost is 67.5 ( 57.2) million with annual costs of 2.7 ( 2.3) million for the UK. Conclusions: Mean egfr was lower with the CKD-EPI formula and individuals were more likely to be reclassified to more advanced CKD. This may be related to the higher mean age of this unselected population compared to previous studies. Refinements of egfr formulae, CKD definitions and cystatin-c eligibility in unselected populations are required. Keywords: chronic kidney disease, egfr, cystatin-c, cost 2

3 Introduction Chronic kidney disease (CKD) is a common condition affecting approximately 7% of the adult population in primary care [1,2]. Glomerular filtration rate is the goldstandard measure of kidney function, but is complex to measure and impractical in everyday clinical practice. Introduced into widespread clinical practice 10 years ago, formulaic estimations of glomerular filtration rate (egfr) were intended to allow more accurate classification of CKD severity compared to serum creatinine alone [3]. Several creatinine-based equations of varying complexity are available for estimating measured glomerular filtration rate, but initially the modification of diet in renal disease (MDRD) egfr equation [4] entered into clinical practice in the UK, supported by CKD guidance from the National Institute for Health and Care Excellence (NICE) [5]. However the alternative CKD-EPI egfr formula [6] may perform better with less bias, especially at higher GFR. It is also superior to the MDRD equation in predicting hard endpoints including time to death, cardiovascular (CV) events and end-stage renal disease (ESRD) [7]. Other egfr equations are based on the serum concentration of other substances, most notably cystatin-c [8]. In an effort to further increase the sensitivity of risk stratification in CKD, NICE CKD guidance was updated and currently recommends the use of the CKD-EPI egfr equation, combined in some circumstances with cystatin-c, for staging of CKD [9]. Specifically, this guidance suggests combining both measures of CKD-EPI and cystatin-c together to confirm or rule out CKD in individuals with potential CKD stage 3a but no proteinuria (albumin:creatinine ratio (ACR) <3mg/mmol) [9]. This recommendation was based on an estimation that the additional healthcare costs related to higher false-positive CKD diagnoses were greater than the additional costs of cystatin-c measurement for this group. However this updated guidance has been controversial and not widely adopted, despite evidence suggesting a dual biochemical approach to CKD classification may better predict patient outcomes [8,10,11]. In England and Wales many laboratories continue to report MDRD egfrs [12] and the uptake of cystatin C egfr into clinical practice has been poor. This inertia may be due to guideline fatigue, concerns 3

4 around workload associated with the implementation of necessary changes by laboratories, and/or costs associated with the resulting need to re-classify some patients to a different CKD stage in primary care. To our knowledge, no study has examined the implications of changing from the MDRD to the CKD-EPI egfr equation for CKD classification in an unselected primary care population, nor quantified the resulting numbers of patients who would subsequently require additional egfr testing by cystatin-c. We therefore used a large UK primary care-based CKD cohort to firstly assess the change in prevalence and classification of CKD if all existing MDRD egfrs were converted to CKD-EPI egfrs; secondly to determine how many patients would fulfil eligibility criteria for an additional cystatin-c egfr test based on prevailing NICE guidance; and thirdly to estimate the initial and ongoing costs associated with this conversion. 4

5 Methods We analysed kidney function data from the Primary-Secondary Care Partnership to Prevent Adverse Outcomes in Chronic Kidney Disease (PSP-CKD) study (ClinicalTrials.gov Identifier: NCT ) cohort [13]. PSP-CKD utilised IMproving Patient care and Awareness of Kidney disease progression Together (IMPAKT), a freely available, web-based CKD audit and quality improvement software tool [14] to support a nurse led CKD management programme in 49 UK general practices. To identify and monitor CKD patients registered at a general practice IMPAKT accesses the practice IT system and uses Morbidity Information Query and Export Syntax (MIQUEST) search methodology to execute queries to identify all adult patients with at least two MDRD egfr values of <60ml/min/1.73m 2, 90 days apart, at any time in the 5 years prior to enrolment of the practice into the study. IMPAKT also extracts other coded CKD-relevant data including anthropometric, demographic, medical history, prescribed medications, blood and urine test results for all patients with CKD stages 3 to 5. Individuals receiving maintenance dialysis or with a renal transplant were excluded. In PSP-CKD data extraction was initially performed at the time of consent and randomisation (baseline, T=0), and then at 6 monthly intervals thereafter for up to 42 months. A pseudonymised dataset from each practice was exported to the University of Leicester Clinical Trials Unit to assemble a prospective database from all participating practices. The study underwent full ethical review from the region s research ethics committee and informed consent was given by the general practices involved in the study. For the current study, serum creatinine values held in the PSP-CKD database were used to calculate both MDRD and CKD-EPI egfrs. At all time-points the relevant clinical chemistry laboratories utilised the Jaffe method for measurement of serum creatinine. egfrs were calculated from serum creatinine data for the T=0 time point, and then for new individuals from each subsequent 6 monthly time point, using standardised formulae [4,6]. 5

6 MDRD: egfr = 175 * (SCr) * (age) * (0.742 if female) * (1.212 if black) CKD-EPI: egfr = 141 * min(scr/κ, 1) α * max(scr/κ, 1) * age * (1.018 if female) * (1.159 if black) For CKD-EPI, κ = 0.7 (females) or 0.9 (males), α = (females) or (males), min indicates the minimum of SCr/κ or 1, and max indicates the maximum of SCr/κ or 1. Serum creatinine values in µmol/l were converted to mg/dl by multiplying by Proteinuria data was derived from urine dipstick results, ACR and protein:creatinine ratio (PCR) values. Results were treated hierarchically with ACR first due to its better correlation to 24 hour urine protein measurement, the gold standard of measurement. Urine dipstick was last in the hierarchy. Individuals were assigned, using egfr and proteinuria data, to a CKD stage based on KDIGO guidelines [15]. Summary cohort data are reported for continuous outcomes as mean ± standard deviation (SD) and for categorical variables as counts and percentage. The t-test was used for comparison of difference in means and net re-classification was calculated based on re-classification to more and less advanced CKD stages. Simple egfr re-categorisation was initially calculated for the study baseline population. Identification and re-classification was then performed for each subsequent time period. MDRD egfr values were plotted against CKD-EPI egfr values for all time points and a linear best fit line was plotted. Patient eligibility for cystatin-c testing was determined based on the criteria described in current NICE CKD guidance for England and Wales [9]. This suggested the use of cystatin-c testing for individuals with a borderline diagnosis of CKD to reduce over-diagnosis. NICE CKD guidance therefore suggested cystatin-c should be considered in individuals with a CKD-EPI egfr of ml/min/1.73 m 2, sustained for at least 90 days and no proteinuria (ACR less than 3 mg/mmol). Cost estimates were based on 2016 standardised national unit costs for health and social care where available [16]. Where not available, cost changes were calculated 6

7 from the NICE 2014 guidance [17] with adjustment for inflation, using the Consumer Price Index, a national UK standard for inflation [18]. All future costs were based on 2017 estimated costs and included calculations of those costs related to taking blood and processing serum cystatin-c tests. The overall costs of cystatin-c were calculated based on these plus the additional related costs of interpretation and potential patient management decisions in primary care. General practitioner consultation times were assumed to be 10 minutes per patient, the same as the 2014 NICE guidance [17]. All statistical analysis, including recalculation of egfrs, was performed using Stata version

8 Results Altogether 49 GP practices, with 353,256 registered patients 18 years of age, were recruited into the PSP-CKD study at baseline. Initially, 21,544 individuals had 2 MDRD egfr results <60 ml/min/1.73m 2 at least 3 months apart. Individuals were eligible to enter the study at each sixth monthly data collection time point and for this analysis a total of 24,660 individuals were included across the study s time period. Table 1 shows selected baseline characteristics of the cohort. At entry into the study 10,031 (46.6%) individuals were categorised by MDRD egfr as CKD stage 3a and proteinuria stage A1 (3a-A1). Table 2 shows the differences in CKD stage classification of the baseline PSP-CKD study population when determined by either the MDRD or CKD-EPI egfr equations. A scatter plot of all MDRD and CKD-EPI egfrs is shown in figure 2 with staging reclassification shown in table 2. At baseline, compared to MDRD egfr, the use of CKD-EPI egfr resulted in re-classification of 2871 (13.3%) CKD patients to a different CKD stage, with 2599 (12.0%) moving to a more advanced CKD stage and 272 (1.3%) moving to a less advanced CKD stage. Two hundred and forty-nine (1.2%) individuals with MDRD stage 3a CKD were re-classified by CKD-EPI to stage 2. The mean change in egfr with conversion to CKD-EPI egfr was ml/min/1.73m2 (95% CI to -2.91, t-test p<0.0001). 9,849 individuals, 39.9% of the CKD population and 2.48% of the adult population, were classified as CKD-EPI stage 3a-A1, thus potentially being eligible for a cystatin- C egfr according to NICE guidance [9]. Since KDIGO CKD guidance differs slightly by suggesting cystatin-c egfr in CKD 3a-A1 where individuals had no other risk factors for CKD (15), we further classified this group by the absence coded diagnoses for DM or HTN. Of these 2,264 individuals (9.2%) had no history of HTN or DM of which 929 (41.0%) were classified as CKD-EPI CKD 3a-A1. Based on similar assumptions to those used by NICE on the economic evaluation of cystatin-c, we estimated the initial and ongoing costs for cystatin-c measurement. The cost of performing a single test of cystatin-c, including phlebotomy costs was 8

9 estimated to be 7.02 ( 5.95). With the addition of primary care physician costs, the total cost estimate was ( 44.95) per patient. From the UK mid-2015 adult population of 51.3 million [19], 1.3 million individuals (2.48% of the adult population) were estimated to be eligible for cystatin-c. When primary care consultation costs [17] were included a cost estimate of 67.5 ( 57.2) million for implementing cystatin- C testing for the eligible UK adult population was calculated. The annual cost for cystatin-c testing in incident CKD 3a-A1 patients was calculated to be 2.7 ( 2.3) million for 50,850 patients (0.1% of the adult population). Table 3 shows the numbers underpinning these calculations. 9

10 Discussion The evolution of egfr formulae has been driven by a need for better prediction of hard outcomes such as ESRF, CV events and mortality [7]. The derivation of both the MDRD and CKD-EPI has been through the collaborative combination of CKD research cohorts [4,6], generally including individuals with well phenotyped CKD, including with heavy proteinuria, and/or of younger age. Whilst KDIGO guidelines have consolidated CKD staging into secondary care [15], national guidelines such as NICE in the UK [9] strongly influence the primary care approach to CKD, where quantitatively the vast majority of care occurs. Therefore the impact of these latter national guidelines is important in allocating healthcare resources. Previous studies have suggested that use of the CKD-EPI formula would reduce the prevalence of CKD and allow focus on individuals at higher risk for renal and CV outcomes [7]. CKD-EPI egfr is more accurate, less bias and more precise than the MDRD formula in estimating measured glomerular filtration rate [6]. For a given egfr, CKD-EPI is also a better predictor of renal, CV and mortality endpoints [7]. Used in isolation cystatin-c and CKD-EPI egfrs have similar levels of bias and precision [20]. However, this is improved with the combination of EPI-CKD and cystatin-c egfr formulae, including in subpopulations of gender, age, level of kidney function and the presence of diabetes mellitus. In CKD, risk prediction for both all-cause and CV mortality is improved with either cystatin-c alone or in combination with CKD-EPI egfr [8]. The first finding of our study suggests that overall in a unselected primary care CKD population, re-classification of CKD stage by CKD-EPI results most commonly in transfer of individuals to a more advanced stage of CKD, with a mean reduction in egfr of around 3 ml/min/1.73m 2. This is in contrast to previous findings where such reclassification led to more patients with less advanced CKD stages [7]. Therefore, this suggests that without the introduction of cystatin-c testing, the prevalence of CKD would arbitrarily rise with the use of the CKD-EPI formula instead of MDRD. The second finding of our study identifies the potential costs of testing a relatively large proportion, in the region of 3-4%, of the adult population to attempt to rule out 10

11 CKD. The current data show that if all eligible CKD-EPI KDIGO stage 3a-A1 patients were to undergo cystatin-c as per prevailing guidance [9], then approximately 2.5% of the adult population would potentially qualify for cystatin-c with a total initial cost to the UK health service of 67.5 ( 57.2) million. This figure does not account for the approximate 25% of the current cohort lacking a proteinuria measurement, and thus may underestimate the true population eligible. Indeed, NHANES data suggests that 3.6% of the adult population may be eligible for cystatin-c measurement under the NICE criteria [15]. On an individual level, combined cystatin-c and creatinine measurement provides incremental improvement in predicting measured glomerular filtration rate and hard outcomes in individuals. It would be associated with substantial initial costs to the healthcare system, even when limited to CKD stage 3a-A1. This therefore brings into question whether cystatin-c measurement in CKD management is the most appropriate use of resources from a public health perspective. The present cohort is probably the largest CKD-specific primary care cohort worldwide, and is derived from a total adult population of approximately 350,000 individuals. The cohort s unselective nature from a large primary care CKD population is one of its major strengths. This representativeness compared to other, mainly secondary care based, CKD studies has allowed the study of the true rates of CKD reclassification from MDRD to CKD-EPI. The difference in reclassification compared to previous studies [7] is mainly explained by a higher mean age in our unselected primary care CKD population (see supplementary material). Indeed, subanalysis of the CKD Prognosis Consortium results suggests that those individuals reclassified to a more advanced CKD category were more likely to have a higher mean age and conversely a lower mean age if reclassified to a less advanced CKD stage [7]. Further, a recent smaller study of 600 older individuals with a median age of 87 years both with and without CKD, suggested a similar reduction of 2.7 ml/min/1.73m 2 in egfr when conversion between MDRD and CKD-EPI was performed [21]. Other studies have also considered the impact of egfr formulae reclassification in older populations compared to measured glomerular filtration rate. The CKD-EPI formula has been shown to be more accurate and less biased than the MDRD 11

12 formula in an older cohort with mean age of 80 years [22]. The more recently developed Berlin Initiative Study (BIS) egfr formulae though may outperform both in older individuals [23]. The present study has a number of limitations. Firstly, as the study population was defined by an MDRD egfr <60 ml/min/1.73m 2 we were unable to assess reclassification from MDRD egfr 60 ml/min/1.73m 2 to CKD-EPI egfr <60 ml/min/1.73m 2. Based on previous literature it is predicted that this classification would be relatively small [7], but the findings of the current study suggest this population may be larger than previously anticipated. This may be another factor that underestimates the true eligibility for cystatin C in the whole adult population. Secondly, the current study is unable to assess the impact of this reclassification on predicting renal and CV outcomes and whether all individuals with CKD 3a-A1 should be considered for cystatin-c measurement or a specific sub-population within this group [19]. Finally, this analysis of cystatin-c includes a large proportion of missing data for proteinuria classification. Again, this is likely to underestimate the potential population requiring cystatin-c measurement. The current results suggest that if individuals with established CKD risk factors such as DM and HTN are assumed to have confirmed CKD and therefore do not require a confirmatory cystatin-c, then only 9.2% of the CKD 3a-A1 population would require cystatin C. The current cost analysis makes a number of assumptions. It was assumed that cystatin-c measurement was taken at a separate time to a second confirmatory egfr measurement. This is because approximately a third of our cohort who had a single egfr <60 ml/min/1.73m2 had a repeat egfr 60 ml/min/1.73m2 and therefore would not require a cystatin-c measurement. Like NICE [9], we were unable to establish a precise unit cost for cystatin-c measurement and therefore the cost estimate was based on that used by NICE with consideration of price inflation. As this was a cross-sectional analysis of egfr measures it was not possible to update any of the NICE s economic assumptions regarding the future impact on management of ruling out CKD for individuals. Whilst the current study cannot provide additional guidance and cost-benefit analysis regarding which subgroups of CKD to test, currently the proportion of the adult populations recommended for cystatin-c measurement is substantial in both number 12

13 and cost. Therefore, a refinement to guidance for public health decision making is required if cystatin-c is to be considered as a tool to rule out CKD as is currently suggested. Targeted cystatin-c measurement may be warranted in those with creatinine based EPI CKD stage 3A-a1 without pre-existing diabetes mellitus or hypertension. Overall, our study suggests that in an unselected primary care CKD cohort, conversion from MDRD egfr to CKD-EPI egfr will increase overall population CKD prevalence bringing more individuals into the earliest stage of CKD. Whether this will practically alter care, or lead to improvements or changes in outcomes for such patients remains unclear. However, this will lead to substantial initial costs of cystatin-c testing for healthcare services, estimated at approximately 67.5 ( 57.2) million in the UK with further on-going annual costs of 2.7 ( 2.3) million. Therefore, further refinement of egfr formulae, CKD definitions and eligibility for cystatin-c at an unselected population based level are required. 13

14 Acknowledgements The authors would like to thank all primary care practices and Nene Clinical Commissioning Group for participating in the PSP-CKD study. Conflict of Interest Statement The authors declare no conflicts of interest. Authors Contributions All authors drafted, edited and agreed the final manuscript for submission. NJB and DS planned and carried out the PSP-CKD study. DS created the IMPAKT tool. RM performed the data analysis for this manuscript. Statements of Ethics The PSP-CKD study was approved by the local research ethics committee prior to it commencing. No additional ethical approval was required for the current analysis. Funding The PSP-CKD study was funded by the National Institute for Health Research (NIHR) Collaboration for Leadership in Applied Health Research and Care (CLAHRC) East Midlands. Ongoing support for the study is funded by NIHR CLAHRC East Midlands and Kidney Research UK (Grant TF2/2015). 14

15 References (1) Bruck K, Stel VS, Gambaro G, Hallan S, Volzke H, Arnlov J, et al. CKD Prevalence Varies across the European General Population. J Am Soc Nephrol 2016 Jul;27(7): (2) Stevens P, O'donoghue D, De Lusignan S, Van Vlymen J, Klebe B, Middleton R, et al. Chronic kidney disease management in the United Kingdom: NEOERICA project results. Kidney Int 2007;72(1): (3) Inker LA, Astor BC, Fox CH, Isakova T, Lash JP, Peralta CA, et al. KDOQI US commentary on the 2012 KDIGO clinical practice guideline for the evaluation and management of CKD. Am J Kidney Dis 2014 May;63(5): (4) Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Ann Intern Med 1999;130(6): (5) Crowe E, Halpin D, Stevens P. Guidelines: early identification and management of chronic kidney disease: summary of NICE guidance. BMJ: British Medical Journal 2008: (6) Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF, Feldman HI, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med 2009;150(9): (7) Matsushita K, Mahmoodi BK, Woodward M, Emberson JR, Jafar TH, Jee SH, et al. Comparison of risk prediction using the CKD-EPI equation and the MDRD study equation for estimated glomerular filtration rate. JAMA 2012;307(18): (8) Shlipak MG, Matsushita K, Ärnlöv J, Inker LA, Katz R, Polkinghorne KR, et al. Cystatin C versus creatinine in determining risk based on kidney function. N Engl J Med 2013;369(10): (9) National Institute for Health and Clinical Excellence. Chronic kidney disease in adults: assessment and management (CG182). 2015; Available at: (10) Shlipak MG, Katz R, Sarnak MJ, Fried LF, Newman AB, Stehman-Breen C, et al. Cystatin C and prognosis for cardiovascular and kidney outcomes in elderly persons without chronic kidney disease. Ann Intern Med 2006;145(4): (11) Stevens LA, Coresh J, Schmid CH, Feldman HI, Froissart M, Kusek J, et al. Estimating GFR using serum cystatin C alone and in combination with serum creatinine: a pooled analysis of 3,418 individuals with CKD. American Journal of Kidney Diseases 2008;51(3):

16 (12) Dr Dorothea Nitsch, Dr Ben Caplin, Dr Sally Hull, Professor David Wheeler. National Chronic Kidney Disease Audit, National Report (Part 1) (13) Major R, Shepherd D, Warwick G, Brunskill N. Prescription Rates of Cardiovascular Medications in a Large UK Primary Care Chronic Kidney Disease Cohort. Nephron 2016;133(1): (14) IMproving Patient care and Awareness of Kidney disease progression Together (IMPAKT). Available at: 17th September (15) Eknoyan G, Lameire N, Eckardt K, Kasiske B, Wheeler D, Levin A, et al. KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int 2013;3:5-14. (16) Curtis L BA. Unit Costs of Health and Social Care (17) National Institute of Health and Clinical Excellence. Chronic kidney disease (partial update) Early identification and management of chronic kidney disease in adults in primary and secondary care Guideline: Appendix L. 2014; Available at: (18) Office for National Statistics. 2017; Available at: (19) Office for National Statistics. Population Estimates for UK, England and Wales, Scotland and Northern Ireland: mid ; Available at: ulationestimates/bulletins/annualmidyearpopulationestimates/latest. Accessed 18/2, (20) Inker LA, Schmid CH, Tighiouart H, Eckfeldt JH, Feldman HI, Greene T, et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med 2012;367(1): (21) Bustos-Guadano F, Martin-Calderon JL, Criado-Alvarez JJ, Munoz-Jara R, Cantalejo-Gutierrez A, Mena-Moreno MC. Glomerular filtration rate estimation in people older than 85: Comparison between CKD-EPI, MDRD-IDMS and BIS1 equations. Nefrologia 2017 Mar - Apr;37(2): (22) Kilbride HS, Stevens PE, Eaglestone G, Knight S, Carter JL, Delaney MP, et al. Accuracy of the MDRD (Modification of Diet in Renal Disease) study and CKD-EPI (CKD Epidemiology Collaboration) equations for estimation of GFR in the elderly. American Journal of Kidney Diseases 2013;61(1): (23) Schaeffner ES, Ebert N, Delanaye P, Frei U, Gaedeke J, Jakob O, et al. Two novel equations to estimate kidney function in persons aged 70 years or older. Ann Intern Med 2012;157(7):

17 Tables Variable n = 24,660 Female 15,265 (61.9) Age 75.2 ± 11.4 MDRD egfr 49.7 (9.2) CKD-EPI egfr 46.8 (9.6) Diabetes Mellitus 5,367 (21.8) CV Disease 4,049 (16.4) PCR (mg/mmol) 37.6 ± 95.8 ACR (mg/mmol) 7.9 ± 26.0 Table 1 Descriptors of whole cohort. Abbreviations ACR albumin creatinine ratio, CV cardiovascular, egfr units ml/min/1.73m², PCR protein creatinine ratio. Results are expressed as mean ± SD, or by frequency (percentage) as appropriate. 17

18 KDIGO Proteinuria Stage No data A1 A2 A3 Total MDRD Stage 3a 4,518 10,031 1, ,529 3b 960 2, , Total 5,730 12,688 2, ,544 CKD-EPI Stage No data A1 A2 A3 Total a 3,871 8,762 1, ,255 3b 1,382 3, , , Total 5,730 12,688 2, ,544 MDRD stage CKD-EPI Stage 2 3a 3b 4 5 Total 3a 249 (1.5) 14,233 (86.1) 2,047 (12.4) 0 (0.0) 0 (0.0) 16,529 3b 0 (0.0) 22 (0.6) 3,468 (86.9) 501 (12.6) 0 (0.0) 3, (0.0) 0 (0.0) 1 (0.1) 890 (94.5) 51 (5.4) (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 82 (100) 82 Total ,255 5,516 1, ,544 Table 2 Classification of CKD at the study s baseline by MDRD, CKD-EPI and proteinuria stage and reclassification of CKD stages between MDRD and CKD-EPI. For reclassification, figures in bold represent individuals not reclassified between MDRD and CKD-EPI. Figures in parentheses are the 18

19 proportion of each MDRD stage classified in each CKD-EPI stage. 19

20 Category T0 (baseline) T1 T2 T3 T4 T5 T6 All Dec 10 June 11 Dec 11 June 12 Dec 12 June 13 Dec 13 n % n % n % n % n % n % n % n % Population searched 353, , ,576^ 372, , , , ,819 MDRD cohort 21, , CKD-EPI cohort 21, , MDRD 3a-A1 10, , CKD-EPI 3a-A1 8, , Estimated UK Adult population* 51,339,161 Estimated population initially eligible for cystatin-c 1,273,393 Estimated incidence of eligibility for cystatin-c 0.10% Estimated initial cystatin-c cost 7,576,689 Estimated ongoing cystatin-c cost per annum 302,556 Estimated total initial costs 57,239,017 Estimated ongoing costs per annum 2,285,698 Table 3 Estimated populations and costs initially and annually eligible for cystatin-c. ^Between T1 and T2, one practice underwent a merger with another practice outside of the study and hence the total population rose by just over 13,000 adults. *Estimated mid-2015 UK adult population from ONS (19). 20