Clinicl Prctice: Mini-Review Received: My 20, 2016 Accepted fter revision: December 14, 2016 Published online: Jnury 27, 2017 Risks for All-Cuse Mortlity: Strtified by Age, Estimted Glomerulr Filtrtion Rte nd Albuminuri Dvid G. Wrnock Pierre Delnye c Richrd J. Glssock b Deprtment of Medicine, University of Albm t Birminghm, Birminghm, AL, nd b Deprtment of Medicine, Geffen School of Medicine t UCLA, Los Angeles, CA, USA; c Nephrology-Dilysis-Trnsplnttion, University of Liège, Liège, Belgium Keywords Serum cretinine Incidence rte rtios Reltive risk Absolute risk Abstrct The reltive reduction of risk for ll-cuse mortlity mong the elderly hs been described s ttenuted compred to younger subjects, but the risk for ll-cuse mortlity increses with ge when bsolute risk is considered. The objectives of the present inquiry were to clculte the reltive nd bsolute risks of ll-cuse mortlity bsed on incidence rte rtios, nd summrize these risks for ge, gender, nd rce strt. We conclude tht strt-specific risk summries re more informtive thn single ll-encompssing risk summry, nd tht gender- nd rce-specific strt need to be considered s well s ge-specific strt. 2017 S. Krger AG, Bsel Introduction Chronic kidney disese (CKD) stging, bsed on estimted glomerulr filtrtion rte (egfr) ctegories, ws recently revised to include risk strt for urinry lbumin-to-cretinine rtio (ACR) [1]. The ssocitions between egfr nd ACR ctegories by ge for ll-cuse mortlity risk nd end-stge renl disese (ESRD) hve been studied in met-nlysis [2]. Our objective ws to compre reltive risks nd bsolute risks for ll-cuse mortlity mong dults strtified by ACR, nd demogrphic ctegories. Previous publictions hve emphsized n ttenution of reltive risk for ll-cuse mortlity for older individuls compred to younger individuls [3 5], nd hve suggested doption of n ge-clibrted definition of CKD. We now report similr pprent ttenution of reltive risk for llcuse mortlity for femle gender nd blck subjects in the study cohort. Rther thn dopting strt-specific definitions of CKD, we believe the sme ultimte gol cn be chieved by dopting strt-specific risk summries ( het mps ) to describe the risk for ll-cuse mortlity. Mterils nd Methods The Resons for Geogrphic nd Rcil Differences in Stroke (REGARDS) study cohort included 30,239 blck nd white individuls who were 45 yers, between Jnury 2003 nd October 2007 [6, 7]. Additionl methodologicl detils nd description of the cohort re provided elsewhere [6 8]. The present nlysis excluded Contribution from the CME course of the DIABESITY Working Group of the ERA-EDTA, Bergmo, December 4 5, 2015. E-Mil krger@krger.com www.krger.com/nef 2017 S. Krger AG, Bsel Prof. Dvid G. Wrnock Deprtment of Medicine, University of Albm t Birminghm Room 614 ZRB, 1720 2nd Avenue South Birminghm, AL 35294-0007 (USA) E-Mil dwrnock @ ubmc.edu Downloded by: Bibliotheque des Sciences de l Vie
Tble 1. Bseline chrcteristics nd outcomes for ge-strt: 26,887 prticipnts with 3,980 deths (15%), excluding prticipnts with missing vlues Chrcteristics Totl Age: 60 Age: 60 70 Age: 70 Ptients, n (%) 26,887 (100) 9,384 (35) 9,215 (34) 8,288 (31) Age, yers, medin (IQR) 64 (58 71) 56 (52 58) 65 (63 67) 75 (72 79) b Mles, % 12,251 (46) 3,836 (41) 4,322 (47) 4,093 (49) b Blcks, % c 10,670 (40) 4,084 (44) 3,689 (40) 2,897 (35) b egfr_cret, ml/min/1.73 m 2, men (SD) 86 (19) 96 (17) 86 (17) 74 (18) b egfr_cret <60 ml/min/1.73 m 2, % 2,760 (10.3) 260 (2.8) 726 (7.9) 1,774 (21) b ACR >30 mg/g, % 3,796 (14) 986 (11) 1,207 (13) 1,603 (19) b SBP, mm Hg, men (SD) 127 (17) 124 (16) 128 (16) 131 (17) b Totl cholesterol, men (SD) 192 (40) 197 (40) 193 (40) 187 (40) b BMI, kg/m 2, medin (IQR) 28 (25 32) 29 (25 34) 29 (25 33) 27 (24 31) b CVD, % 5,692 (21) 1,192 (13) 1,944 (21) 2,556 (31) b Dibetes, % 5,413 (20) 1,627 (17) 2,026 (22) 1,760 (21) Current smoking, % 3,820 (14) 1,860 (20) 1,315 (14) 645 (7.8) b Outcomes, % Deths, % 3,980 (15) 505 (5.4) 1,032 (11) 2,443 (30) b Incidence rte, per 1,000-person yers 21.3 (20.7 22.0) 7.7 (7.1 8.5) 15.5 (14.6 16.5) 45 (43 46) b Incidence rte rtio 1.00 (reference) 2.00 (1.80 2.22) 5.74 (5.22 6.32) b Vlues re given s number (percentge), mens (SD); medins (IQR); incidence rtes nd incidence rte rtios (95% CI). ACR, urinry lbumin-to-cretinine rtio; BMI, body mss index; CVD, prevlent crdiovsculr disese; egfr_cret, cretininebsed estimted glomerulr filtrtion rte, CKD Epidemiology Collbortion; IQR, interqurtile rnge, 25th 75th qurtile; SBP, systolic blood pressure. p < 0.001, bold text used to indicte significnce for comprisons of chrcteristics to the reference group by ANOVA for mens, signed rnk test for medins, nd chi-squre test for proportions. Row percentges; ll others re column percentges. b p < 0.05 for comprison between 3rd nd 2nd ctegories (Dunnett test for multiple comprisons) or chi-squre test or proportions. c Rce or ethnic group ws self-reported, nd further clssified s blck or non-blck. prticipnts with missing vlues for covrites. The effect vrible ws the stndrdized cretinine-bsed CKD-Epidemiology Collbortion egfr [9]. All-cuse mortlity ws scertined s previously described [6 8]. The UAB Institutionl Review Bord pproved the protocol nd ll prticipnts gve written informed consent. Mortlity Rtes, Incidence Rtes nd Absolute Rtes for All-Cuse Mortlity Mortlity rtes (percentge), incidence rtes (per 1,000-person yers; Stt commnd stptime), nd incidence rte rtios with ±95% CIs were clculted with multi-vrible djusted Poisson regression models for egfr strt (>100, 80 100, 60 80, 45 60, nd <45 ml/min/1.73 m 2 ), ACR strt (<30 nd 30 mg/g), nd ge strt (<60, 60 70, nd >70 yers). Absolute incidence rtes were obtined from the incidence rte rtios by multiplying ech by the rte for egfr for 80 100 ml/min/1.73 m 2, ACR <30 mg/g, nd ge <60 yers [10]. Dunnett method ws used for mking comprisons with reference ctegory for pirwise comprisons of mens. Incidence rte rtios were clculted with Poisson regression for egfr deciles between 25 nd 125 ml/min/1.73 m 2. Men vlues (±95% CI) re used s the reltive risk for ech intervl with covrites fixed t their men vlues (Stt commnd mrgins, tmens). Stt (version 14.1, College Prk Sttion, TX, USA) ws used for sttisticl nlysis. Results Bseline Chrcteristics Bseline chrcteristics nd deths re shown for the ge strt in Tble 1 for 26,887 with 3,989 deths (15%). Of the entire cohort, 40% were Africn Americn nd 25% were <58 yers of ge t enrollment. Mle prepondernce, systolic blood pressure, prevlent crdiovsculr disese, lbuminuri >30 mg/dl, nd egfr <60 ml/min/1.73 m 2 incresed cross the ge strt. The number of blcks, totl cholesterol, nd current smoking sttus decresed cross the ge strt. As expected, ll-cuse mortlity rtes incresed with ge. Risk Summries for All-Cuse Mortlity: Het Mps Tble 2 shows ll-cuse mortlity cross ge, nd ACR strt, with the number of deths expressed s percentge. Higher mortlity rtes were ssocited with lower egfr ctegories, nd especilly for subjects with Demogrphics, CKD nd All-Cuse Mortlity 293 Downloded by: Bibliotheque des Sciences de l Vie
Tble 2. Counts nd ll-cuse mortlity rtes: ge, nd ACR strt egfr_cret, Age: <60 Age: 60 70 Age: >70 >100 3,370 439 1,430 241 370 80 151 (4.5) 50 (11) 146 (10) 46 (19) 67 (18) 28 (35) 80 100 3,825 309 3,897 418 2,667 480 141 (3.7) 31 (10) 315 (8.1) 86 (21) 597 (22) 178 (37) 60 80 1,037 144 2,189 314 2,437 480 64 (6.2) 19 (13) 200 (9.1) 65 (21) 606 (25) 206 (43) 45 60 145 51 396 137 893 319 15 (10) 11 (22) 49 (12) 44 (32) 288 (32) 163 (51) <45 21 43 96 97 318 244 3 (14) 20 (47) 20 (21) 61 (63) 139 (44) 171 (70) Vlues re n or n (%). The mortlity rtes were clculted s the percentge of deths in ech cell. Color version vilble online b. Reltive risk (±95% CI) for ll-cuse mortlity: ge, nd ACR strt b Age: <60 Age: 60 70 Age: >70 >100 1.20 (0.95 1.52) 2.49 (1.77 3.49) 1.12 (0.97 1.38) 1.78 (1.29 2.45) 0.87 (0.67 1.12) 1.57 (1.07 2.31) 80 100 1.00 (reference) 2.02 (1.36 3.01) 1.00 (reference) 2.28 (1.79 2.91) 1.00 (reference) 1.69 (1.42 2.00) 60 80 1.58 (1.17 2.12) 2.28 (1.40 3.73) 1.12 (0.94 1.34) 1.92 (1.46 2.53) 1.13 (1.01 1.26) 1.88 (1.60 2.12) 45 60 2.11 (1.23 3.61) 4.00 (2.13 7.51) 1.47 (1.08 1.99) 3.21 (2.32 4.43) 1.48 (1.28 1.70) 2.38 (2.00 2.84) <45 2.96 (0.94 9.30) 6.80 (4.08 11.3) 2.59 (1.64 4.08) 7.13 (5.36 9.50) 2.22 (1.84 2.67) 3.71 (3.12 4.42) Incidence rte rtio; reltive to reference group for ech ge strt. Color coding for reltive risk of ll-cuse mortlity: green, <1.5; yellow, >1.5 nd <2.5; ornge, >2.5 nd <4.0; red, >4.0. Color version vilble online. c. Absolute incidence rte (±95% CI) for ll-cuse mortlity: ge, nd ACR strt b Age: <60 Age: 60 70 Age: >70 >100 1.23 (0.05 2.42) 5.37 (3.71 7.04) 4.25 (3.05 5.45) 6.99 (5.27 8.71) 7.29 (5.78 8.80) 10.1 (8.02 12.2) 80 100 5.13 (4.35 6.05)* 4.25 (2.26 6.26) 3.32 (2.30 4.34) 7.95 (6.57 934) 8.40 (7.45 9.35) 10.9 (9.71 12.0) 60 80 2.53 (1.02 4.05) 5.30 (2.82 7.77) 3.94 (2.83 5.05) 7.32 (5.80 8.84) 8.99 (8.04 9.94) 11.3 (10.2 12.4) 45 60 4.89 (2.15 7.62) 8.21 (5.05 11.4) 5.46 (3.79 7.13) 9.93 (8.14 11.7) 10.3 (9.34 11.3) 12.5 (11.3 13.7) <45 6.32 (0.46 12.2) 11.5 (8.96 14.0) 8.47 (6.06 10.9) 14.1 (12.6 15.7) 12.4 (11.1 13.6) 14.8 (13.6-16.0) Incidence rte; deths per 1,000 ptient yer. The incidence rte rtios in Tble 2c were clculted for ech gender nd ge >70 yers; the reference group hd egfr >80 nd <100 ml/min/1.73 m 2 nd ACR <30 mg/g cretinine. Color coding for bsolute risk of ll-cuse mortlity: green, <6.0; yellow, >6.0 nd <8.0; ornge, >8.0 nd <12.0; red, >12.0. Color version vilble online. The bsolute mortlity rtes (Tble 3b) for ech cell were obtined by multiply the incidence rte rtios by the mortlity rte for the egfr 80 100, ACR <30 mg/g group in column 1 (indicted by*). ACR, urinry lbumin-to-cretinine rtio; cretinine-bsed estimted glomerulr filtrtion rte. 294 Wrnock/Delnye/Glssock Downloded by: Bibliotheque des Sciences de l Vie
Tble 3. Reltive risk (±95% CI) for ll-cuse mortlity: rce, ge >70, nd ACR strt Age: >70 Age: >70; white Age: >70; blck >100 0.87 (0.67 1.12) 1.57 (1.07 2.31) 1.67 (0.86 3.24) 9.09 (2.91 28.4) 0.80 (0.60 1.08) 1.44 (0.94 2.19) 80 100 1.00 (reference) 1.69 (1.42 2.00) 1.00 (reference) 1.83 (1.49 2.24) 1.00 (reference) 1.45 (1.08 1.96) 60 80 1.13 (1.01 1.26) 1.88 (1.60 2.12) 1.10 (0.96 1.26) 1.87 (1.53 2.28) 1.17 (0.95 1.44) 1.90 (1.44 2.50) 45 60 1.48 (1.28 1.70) 2.38 (2.00 2.84) 1.54 (1.30 1.81) 2.34 (1.88 2.82) 1.34 (1.03 1.75) 2.42 (1.80 3.26) <45 2.22 (1.84 2.67) 3.71 (3.12 4.42) 2.14 (1.71 2.69) 3.21 (2.54 4.06) 2.30 (1.65 3.21) 4.58 (3.50 6.00) Color version vilble online b. Reltive risk for ll-cuse mortlity: gender, ge >70, nd ACR strt Age: >70 Age: >70; mles Age: >70; femles >100 0.87 (0.67 1.12) 1.57 (1.07 2.31) 1.27 (0.91 1.77) 1.25 (0.70 2.25) 0.57 (0.38 0.87) 1.95 (1.17 3.27) 80 100 1.00 (reference) 1.69 (1.42 2.00) 1.00 (reference) 1.72 (1.38 2.14) 1.00 (reference) 1.68 (1.29 2.20) 60 80 1.13 (1.01 1.26) 1.88 (1.60 2.12) 1.21 (1.05 1.40) 1.91 (1.55 2.34) 1.00 (0.83 1.20) 1.90 (1.46 2.46) 45 60 1.48 (1.28 1.70) 2.38 (2.00 2.84) 1.62 (1.34 1.95) 2.54 (2.04 3.17) 1.30 (1.05 1.62) 2.15 (1.61 2.88) <45 2.22 (1.84 2.67) 3.71 (3.12 4.42) 2.57 (1.98 3.35) 3.47 (2.75 4.38) 1.87 (1.43 2.43) 4.12 (3.18 5.35) Incidence rte rtio; reltive to the reference group for ech ge strt. Color coding for bsolute risk of ll-cuse mortlity: light green, <1.0; green, >1 nd <1.5; yellow, >1.5 nd <2.5; ornge, >2.5 nd <4.0; red, >4.0. Color version vilble online. ACR, urinry lbumin-to-cretinine rtio; cretinine-bsed estimted glomerulr filtrtion rte [9]. ACR >30 mg/g. As expected, higher mortlity rtes were observed with incresing ge ctegories for ll strt. Reltive risks for mortlity cross the ge strt, reltive to common reference t egfr between 80 nd 100 ml/min/1.73 m 2 nd ACR <30 mg/g re shown in Tble 2 b. Arbitrry color coding corresponds to incresing reltive risk (see legend). The reltive risk for ll-cuse mortlity ws ttenuted for older subjects, nd the reltive risk is similr for egfr strt >45 ml/min/1.73 m 2 nd ACR <30 mg/g. Tble 2 c shows the bsolute risk for mortlity cross the ge strt. Incidence rte rtios were clculted reltive to common reference t egfr between 80 nd 100 ml/min/1.73 m 2, ACR <30 mg/g, nd ge <60 yers. Arbitrry color coding corresponds to incresing bsolute risk (see legend). The bsolute risk for llcuse mortlity ws incresed for older REGARDS prticipnts (ge >70 yers), nd the bsolute risk ws similr for egfr strt >45 ml/min/1.73 m 2, nd ACR <30 mg/g ws similr for subjects <70 yers. The bsolute risk for mortlity incresed with ge nd ACR >30 mg/g. Tble 3 shows the reltive risk for mortlity for rce nd gender strt, mong subjects >70 yers of ge. Arbitrry color coding corresponds to incresing bsolute risk (see legend). The reltive risk for ll-cuse mortlity ws ttenuted for blck subjects compred to white subjects >70 yers of ge ( Tble 3 ), nd for femle subjects compred to mle subjects >70 yers of ge nd ACR <30 mg/g ( Tble 3 b). Discussion We describe mortlity risk for REGARDS prticipnts bsed on demogrphic strt, nd ACR ctegories using incidence rtes nd incidence rte rtios for llcuse mortlity. Risk summries ( het mps ) show notble differences cross demogrphic strt. The reltive risks for ll-cuse mortlity were ttenuted for the older ge strt ( Tble 2 b), while the bsolute risk for mortlity incresed with ge ( Tble 2 c). For exmple, the reltive risks for mortlity mong prticipnts with egfr >60 nd <80 ml/min/1.73 m 2, nd Demogrphics, CKD nd All-Cuse Mortlity 295 Downloded by: Bibliotheque des Sciences de l Vie
ACR <30 mg/g ws 1.58 (95% CI 1.17 2.12) for ge <60 yers; 1.12 (95% CI 0.94 1.34) for ge >60 nd <70 yers; nd 1.13 (95% CI 1.01 1.26) for ge >70 yers ( Tble 2 b). In contrst, the bsolute risk for mortlity mong the sme prticipnts ws incresed cross the ge strt: 2.53 (95% CI 1.02 4.05) for ge <60 yers; 3.94 (95% CI 2.83 5.05) for ge >60 nd <70 yers; nd 8.99 (95% CI 8.04 9.94) for ge >70 yers ( Tble 2 c). Similr ptterns were seen cross the sme egfr nd ge strt for ACR >30 mg/g. Previous commentries hve emphsized the ttenution of the reltive risk for ll-cuse mortlity mong older subjects [4]. The present nlysis confirms this finding, but lso extends the nlysis to show the expected increse in bsolute risk for ll-cuse mortlity with incresing ge. Similr observtions bout the increse in mortlity rtes with ge nd ACR levels re described by Hlln et l. [2]. Furthermore, the gender nd rce differences in reltive risk for ll-cuse mortlity shown in Tble 3 nd b suggest tht these strt s well s ge should lso be considered when ssessing the reltive risk for ll-cuse mortlity mong specified cohort of ptients. This conclusion is consistent with longitudinl descriptions of the risk for ll-cuse mortlity described for elderly Swedish women [11]. The KDIGO Controversies Conference Report [10] included het mps for seprte outcome events, including ll-cuse mortlity, crdiovsculr mortlity, ESRD, cute kidney injury, nd progressive CKD. The reltive risk summry tble for ll-cuse mortlity [10] is similr to the bsolute risk summry. It used similr reference groups (ACR <10 nd <30 mg/g, egfr 90 105 nd 80 100 ml/min/1.73 m 2 ), with mortlity incidence rtes of 7.0 nd 5.13 deths per 1,000-person yers, respectively. Absolute risk in Tble 2 c ws generted by multiplying incidence rte rtios by the bsolute risk for the reference group, s previously described [10]. As expected, the risk for ll-cuse mortlity incresed with ging. With the gol of confirming single ll-encompssing risk summry scheme [1], Hlln et l. [2] exmined risk summries for ge strt, nd ACR ctegories with met-nlysis of 33 generl popultion cohorts, including over 2 million prticipnts with 112,325 deths, including 27,306 REGARDS prticipnts. Risk for ll-cuse mortlity differ cross ge strt, both for reltive risk nd bsolute risk [2], supporting the use of different risk summries for different outcome events, demogrphic strt, nd egfr nd ACR strt. We conclude tht the differences between reltive nd bsolute risks for ll-cuse mortlity hve not received sufficient emphsis. Furthermore, there my be differences in the bsolute nd reltive risks cross ge-strt tht require further nlyses for gender nd rce strt. The REGARDS cohort is not lrge enough to fully define these strt, nd ws recruited with the gol of including subjects >45 yers of ge, nd n equl number of men, women, nd white nd blck prticipnts. Additionl met-nlysis of combined generl popultion cohorts seems worthwhile. It hs been proposed tht the cutpoints for CKD stges be djusted for ge strt, described s n ge-clibrted definition of CKD [3 5], to ccount for the ttenution of reltive risk for ll-cuse mortlity observed with older subjects. This pproch hs not been widely ccepted [12], nd represents post-hoc modifiction of risk ssessment tht does not ddress the underlying weknesses inherent when such nlyses include egfr s n effect vrible [8]. The current proposl to use seprte het mps for ge, rce, nd gender strt rther thn single ll-encompssing risk summry my ccomplish the bsic gol of providing risk ssessments in CKD tht reflect the demogrphic nd kidney-specific fctors for n individul ptient. Acknowledgments This reserch project is supported by coopertive greement U01 NS041588 from the Ntionl Institute of Neurologicl Disorders nd Stroke, Ntionl Institutes of Helth, Deprtment of Helth nd Humn Service. The content is solely the responsibility of the uthors nd does not necessrily represent the officil views of the Ntionl Institute of Neurologicl Disorders nd Stroke or the Ntionl Institutes of Helth. Representtives of the funding gency hve been involved in the review of the mnuscript but not directly involved in the collection, mngement, nlysis or interprettion of the dt. The uthors thnk the other investigtors, the stff, nd the prticipnts of the REGARDS study for their vluble contributions. A full list of prticipting REGARDS investigtors nd institutions cn be found t http://www.regrdsstudy.org. The support of the UAB/UCSD O Brien Center for Kidney Reserch (P30 DK079337) is cknowledged. Additionl funding ws provided by n investigtor-initited grnt-in-id from Amgen Corportion. Amgen did not hve ny role in the design nd conduct of the study or the preprtion of the mnuscript. Disclosure Sttement The uthors hve no conflicts of interest to declre. 296 Wrnock/Delnye/Glssock Downloded by: Bibliotheque des Sciences de l Vie
References 1 Kidney Disese: Improving Globl Outcomes (KDIGO) CKD Work Group: KDIGO 2012 clinicl prctice guideline for the evlution nd mngement of chronic kidney disese. Kidney Int Suppl 2013; 3: 1 150. 2 Hlln SI, Mtsushit K, Sng Y, Mhmoodi BK, Blck C, Ishni A, Kleefstr N, Nimrk D, Roderick P, Tonelli M, Wetzels JF, Astor BC, Gnsevoort RT, Levin A, Wen CP, Coresh J; Chronic Kidney Disese Prognosis Consortium: Age nd ssocition of kidney mesures with mortlity nd end-stge renl disese. JAMA 2012; 308: 2349 2360. 3 Glssock R, Delnye P, El Nhs M: An geclibrted clssifiction of chronic kidney disese. JAMA 2015; 314: 559 560. 4 Delnye P, Glssock RJ, Pottel H, Rule AD: An ge-clibrted definition of chronic kidney disese: rtionle nd benefits. Clin Biochem Rev 2016; 37: 17 26. 5 Denic A, Glssock RJ, Rule AD: Structurl nd functionl chnges with the ging kidney. Adv Chronic Kidney Dis 2016; 23: 19 28. 6 Howrd VJ, Cushmn M, Pulley L, Gomez CR, Go RC, Prines RJ, Grhm A, Moy CS, Howrd G: The resons for geogrphic nd rcil differences in stroke study: objectives nd design. Neuroepidemiology 2005; 25: 135 143. 7 Howrd VJ, Kleindorfer DO, Judd SE, Mc- Clure LA, Sfford MM, Rhodes JD, Cushmn M, Moy CS, Solimn EZ, Kissel BM, Howrd G: Disprities in stroke incidence contributing to disprities in stroke mortlity. Ann Neurol 2011; 69: 619 627. 8 Wrnock DG: Estimted glomerulr filtrtion rte: fit for wht purpose? Nephron 2016; 134: 43 49. 9 Inker LA, Schmid CH, Tighiourt H, Eckfeldt JH, Feldmn HI, Greene T, Kusek JW, Mnzi J, Vn Lente F, Zhng YL, Coresh J, Levey AS; CKD-EPI Investigtors: Estimting glomerulr filtrtion rte from serum cretinine nd cysttin C. N Engl J Med 2012; 367: 20 29. 10 Levey AS, de Jong PE, Coresh J, El Nhs M, Astor BC, Mtsushit K, Gnsevoort RT, Ksiske BL, Eckrdt KU: The definition, clssifiction, nd prognosis of chronic kidney disese: KDIGO controversies conference report. Kidney Int 2011; 80: 17 28. 11 Mlmgren L, McGuign FE, Berglundh S, Westmn K, Christensson A, Akesson K: Declining estimted glomerulr filtrtion rte nd its ssocition with mortlity nd comorbidity over 10 yers in elderly women. Nephron 2015; 130: 245 255. 12 Levey AS, Inker LA, Coresh J: Chronic kidney disese in older people. JAMA 2015; 314: 557 558. Demogrphics, CKD nd All-Cuse Mortlity 297 Downloded by: Bibliotheque des Sciences de l Vie