Lower estimated glomerular filtration rate and higher albuminuria are associated with all-cause and cardiovascular mortality. A collaborative meta-analysis of high-risk population cohorts

Screening for chronic kidney disease is recommended in people at high risk, but data on the independent and combined associations of estimated glomerular filtration rate (eGFR) and albuminuria with all-cause and cardiovascular mortality are limited. To clarify this, we performed a collaborative meta-analysis of 10 cohorts with 266,975 patients selected because of increased risk for chronic kidney disease, defined as a history of hypertension, diabetes, or cardiovascular disease. Risk for all-cause mortality was not associated with eGFR between 60–105 ml/min per 1.73 m2, but increased at lower levels. Hazard ratios at eGFRs of 60, 45, and 15 ml/min per 1.73 m2 were 1.03, 1.38 and 3.11, respectively, compared to an eGFR of 95, after adjustment for albuminuria and cardiovascular risk factors. Log albuminuria was linearly associated with log risk for all-cause mortality without thresholds. Adjusted hazard ratios at albumin-to-creatinine ratios of 10, 30 and 300 mg/g were 1.08, 1.38, and 2.16, respectively compared to a ratio of five. Albuminuria and eGFR were multiplicatively associated with all-cause mortality, without evidence for interaction. Similar associations were observed for cardiovascular mortality. Findings in cohorts with dipstick data were generally comparable to those in cohorts measuring albumin-to-creatinine ratios. Thus, lower eGFR and higher albuminuria are risk factors for all-cause and cardiovascular mortality in high-risk populations, independent of each other and of cardiovascular risk factors. Screening for chronic kidney disease is recommended in people at high risk, but data on the independent and combined associations of estimated glomerular filtration rate (eGFR) and albuminuria with all-cause and cardiovascular mortality are limited. To clarify this, we performed a collaborative meta-analysis of 10 cohorts with 266,975 patients selected because of increased risk for chronic kidney disease, defined as a history of hypertension, diabetes, or cardiovascular disease. Risk for all-cause mortality was not associated with eGFR between 60–105 ml/min per 1.73 m2, but increased at lower levels. Hazard ratios at eGFRs of 60, 45, and 15 ml/min per 1.73 m2 were 1.03, 1.38 and 3.11, respectively, compared to an eGFR of 95, after adjustment for albuminuria and cardiovascular risk factors. Log albuminuria was linearly associated with log risk for all-cause mortality without thresholds. Adjusted hazard ratios at albumin-to-creatinine ratios of 10, 30 and 300 mg/g were 1.08, 1.38, and 2.16, respectively compared to a ratio of five. Albuminuria and eGFR were multiplicatively associated with all-cause mortality, without evidence for interaction. Similar associations were observed for cardiovascular mortality. Findings in cohorts with dipstick data were generally comparable to those in cohorts measuring albumin-to-creatinine ratios. Thus, lower eGFR and higher albuminuria are risk factors for all-cause and cardiovascular mortality in high-risk populations, independent of each other and of cardiovascular risk factors. The definition and classification of chronic kidney disease was proposed by Kidney Disease Outcomes Quality Initiative (KDOQI) in 2002, and endorsed by Kidney Disease Improving Global Outcomes (KDIGO) in 2004.1.National Kidney Foundation K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification.Am J Kidney Dis. 2002; 39: S1-S266Abstract Full Text Full Text PDF PubMed Scopus (228) Google Scholar, 2.Levey A.S. Coresh J. Balk E. et al.NKF Practice Guidelines for CKD: evaluation, classification and stratification.Arch Int Med. 2003; 139: 137-147Google Scholar, 3.Levey A.S. Eckardt K.U. Tsukamoto Y. et al.Definition and classification of chronic kidney disease: a position statement from Kidney Disease: Improving Global Outcomes (KDIGO).Kidney Int. 2005; 67: 2089-2100Abstract Full Text Full Text PDF PubMed Scopus (2202) Google Scholar Widespread implementation of the definition and classification has promoted increased attention to chronic kidney disease in clinical practice, research, and public health.4.Levey A.S. Atkins R. Coresh J. et al.Chronic kidney disease as a global public health problem: approaches and initiatives – a position statement from Kidney Disease Improving Global Outcomes.Kidney Int. 2007; 72: 247-259Abstract Full Text Full Text PDF PubMed Scopus (947) Google Scholar It has also generated substantial debate about the appropriateness of recommending the same glomerular filtration rate (GFR) thresholds for people of all ages, the optimal level of albuminuria as a marker of kidney damage, and about the value of the 5-stage classification system based on estimated GFR (eGFR) without consideration of albuminuria.5.Gansevoort R.T. de Jong P.E. The case for using albuminuria in staging chronic kidney disease.J Am Soc Nephrol. 2009; 20: 465-468Crossref PubMed Scopus (83) Google Scholar, 6.Glassock R.J. Winearls C. An epidemic of chronic kidney disease: fact or fiction?.Nephrol Dial Transpl. 2008; 23: 1117-1123Crossref PubMed Scopus (169) Google Scholar, 7.Ikizler T.A. CKD classification: time to move beyond KDOQI.J Am Soc Nephrol. 2009; 20: 929-930Crossref PubMed Scopus (26) Google Scholar, 8.Winearls C.G. Glassock R.J. Dissecting and refining the staging of chronic kidney disease.Kidney Int. 2009; 75: 1009-1014Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar In October 2009, KDIGO sponsored a Controversies Conference to examine the validity of the existing system as well as to evaluate proposed alternatives.9.Eckardt K.U. Berns J.S. Rocco M.V. et al.Definition and classification of CKD: the debate should be about patient prognosis – a position statement from KDOQI and KDIGO.Am J Kidney Dis. 2009; 53: 915-920Abstract Full Text Full Text PDF PubMed Scopus (165) Google Scholar The report of the Consensus Conference is included in this issue of Kidney International.10.Levey A.S. de Jong P.E. Coresh J. et al.The definition, classification and prognosis of chronic kidney disease: a KDIGO controversies Conference report.Kidney Int. 8 December 2010; (e-pub ahead of print)Google Scholar As part of the process, the CKD Prognosis Collaboration was formed to undertake a comprehensive analysis of mortality and kidney outcomes according to estimated GFR and albuminuria, to answer key questions underlying the debate. This paper is the second in a series of four papers to report the results of collaborative meta-analyses undertaken by the CKD Prognosis Consortium. The first paper in this series deals with all-cause and cardiovascular mortality in general population cohorts.11.The Chronic Kidney Disease Prognosis Consortium Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality: a collaborative meta-analysis of general and high risk population cohorts.Lancet. 2010; 375: 2073-2081Abstract Full Text Full Text PDF PubMed Scopus (2460) Google Scholar The present report describes all-cause and cardiovascular mortality in cohorts at high risk for chronic kidney disease. Other manuscripts report kidney outcomes from general population and high-risk cohorts,12The Chronic Kidney Disease Prognosis Consortium. Association of estimated glomerular filtration rate and albuminuria with kidney outcomes: a collaborative meta-analysis of general and high risk population cohorts. Kidney Int (submitted).Google Scholar and mortality and kidney outcomes in chronic kidney disease cohorts.13The Chronic Kidney Disease Prognosis Consortium. Association of estimated glomerular filtration rate and albuminuria with mortality and end-stage renal disease: a collaborative meta-analysis of kidney disease cohorts. Kidney Int (submitted).Google Scholar Chronic kidney disease is now recognized as a risk factor for all-cause and cardiovascular mortality.14.Sarnak M.J. Levey A.S. Schoolwerth A.C. et al.American Heart Association Councils on Kidney in Cardiovascular Disease, High Blood Pressure Research, Clinical Cardiology, and Epidemiology and Prevention. Kidney disease as a risk factor for development of cardiovascular disease: a statement from the American Heart Association Councils on Kidney in Cardiovascular Disease, High Blood Pressure Research, Clinical Cardiology, and Epidemiology and Prevention.Circulation. 2003; 108: 2154-2169Crossref PubMed Scopus (2675) Google Scholar Our meta-analysis of 21 general population cohorts showed the independent and joint associations of reduced estimated GFR and higher levels of albuminuria on these outcomes.11.The Chronic Kidney Disease Prognosis Consortium Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality: a collaborative meta-analysis of general and high risk population cohorts.Lancet. 2010; 375: 2073-2081Abstract Full Text Full Text PDF PubMed Scopus (2460) Google ScholarPresently, screening for chronic kidney disease is recommended in people at high risk for chronic kidney disease including patients with cardiovascular disease risk factors.4.Levey A.S. Atkins R. Coresh J. et al.Chronic kidney disease as a global public health problem: approaches and initiatives – a position statement from Kidney Disease Improving Global Outcomes.Kidney Int. 2007; 72: 247-259Abstract Full Text Full Text PDF PubMed Scopus (947) Google Scholar However, the associations between eGFR and albuminuria with mortality may differ in high-risk populations as compared with the general population. As a comprehensive analysis of the associations between eGFR and albuminuria with all-cause and cardiovascular mortality in high-risk populations has not been reported, we studied these associations in a collaborative meta-analysis. A priori we hypothesized that both eGFR and albuminuria would be associated with these outcomes, independent of traditional cardiovascular risk factors and independent of each other, and despite inclusion of diverse study populations. Of 10 high-risk cohorts (266,975 subjects), 6 had data on albumin-to-creatinine ratio (117,500 subjects) and 4 on dipstick proteinuria (149,475 subjects) (Table 1). Acronyms and abbreviations for studies included in the current report are given in Supplementary Table S1 online. By definition, the study participants have a high prevalence of cardiovascular disease risk factors. The characteristics of the dipstick cohorts are, in general, comparable to those of the albumin-to-creatinine ratio cohorts, with the albumin-to-creatinine ratio cohorts having a higher percentage of males and subjects with diabetes and a history of cardiovascular disease, and a lower percentage of Blacks. In the cohorts with information on albumin-to-creatinine ratio, there were 8706 all-cause deaths and 3171 cardiovascular disease deaths during follow-up. In the subjects with dipstick data, these figures are 7303 and 2485, respectively. A total of 36.7% of the subjects in the pooled study population with measurements of albumin-to-creatinine ratio had chronic kidney disease according to the current definition (eGFR <60 ml/min per 1.73 m2 or albumin-to-creatinine ratio ≥30 mg/g) (Supplementary Tables S2 and S3 online). This subgroup accounted for 58.6% of all-cause mortality events (Supplementary Table S4 online) and 59.4% of cardiovascular mortality events (Supplementary Table S5 online). Data on cardiovascular mortality were available in only two of the dipstick cohorts, which differed greatly in sample size and duration of follow-up.Table 1Characteristics of included studiesNAge (year)Male (%)Black (%)CVD (%)HT (%)HC (%)DM (%)Smoking (%)eGFR (ml/min per 1.73 m2)ACR (mg/g)FU (Year)ACM (n)CVM (N)Cohorts with ACR data ADVANCE11,14065.857.5NA32.282.233.010015.178.215.94.81031542 AKDN67,40655.856.8NA5.046.8NA49.0NA76.811.12.32371— ONTARGET25,62066.473.32.592NA*NA*37.512.673.652.24.530681821 Pima634126.445.40NA12.94.220.427.814.411.913.51083170 TRANSCEND592666.9571.892.5NA*NA*35.79.871.725.34.6713450 ZODIAC106767.943.4034.983.34010018.663.819.67.8440188Total117,50087063171Weighted mean58.059.21.932.749.323.649.614.979.521.43.8Cohorts with dipstick data CARE409858.686.23.210082.979.014.216.171.9—4.8371211 KEEP92,31654.831.731.313.057.323.030.511.181.2—2.4568— KP Hawaii40,21059.050.4NA17.0NANA48.013.671.5—2.41706— MRFIT12,85146.21007.20.062.357.13.163.779.7—21.646582274Total149,47573032485Weighted mean55.344.127.415.358.829.132.416.478.2—4.1Abbreviations: ACM, all-cause mortality; ACR, albumin-to-creatinine ratio; CVD, cardiovascular disease; CVM, cardiovascular mortality; DM, diabetes mellitus; eGFR, estimated glomerular filtration rate; FU, duration of follow-up; HC, hypercholesterolemia; HT, hypertension; NA, not available.NA* in ONTARGET and TRANSCEND, respectively, a history of hypertension was reported by 69 and 76%, and statin use by 62 and 55%. Open table in a new tab Download .doc (.67 MB) Help with doc files Supplementary Information Abbreviations: ACM, all-cause mortality; ACR, albumin-to-creatinine ratio; CVD, cardiovascular disease; CVM, cardiovascular mortality; DM, diabetes mellitus; eGFR, estimated glomerular filtration rate; FU, duration of follow-up; HC, hypercholesterolemia; HT, hypertension; NA, not available. NA* in ONTARGET and TRANSCEND, respectively, a history of hypertension was reported by 69 and 76%, and statin use by 62 and 55%. Pooled adjusted hazard ratios of all-cause mortality and cardiovascular mortality according to eGFR and albumin-to-creatinine ratio are shown in Figure 1. The association between eGFR and relative risk for all-cause mortality and cardiovascular mortality was relatively constant between 60 and 105 ml/min per 1.73 m2, and steadily increased at eGFR below 60 ml/min per 1.73 m2 (Figure 1, left panels). For all-cause mortality, adjusted hazard ratios at eGFR 60, 45, and 15 ml/min per 1.73 m2 were 1.03 (0.81–1.33), 1.38 (1.15–1.65), and 3.11 (2.26–4.27), respectively, whereas for cardiovascular mortality, the adjusted hazard ratios were 1.11 (0.93–1.32), 1.73 (1.49–2.00), and 3.08 (1.89–5.01), respectively. The pattern for all-cause mortality was comparable in the dipstick cohorts (Supplementary Figure S1 online), except there was a tendency for a U-shaped relationship with significantly decreased risk at eGFR 75 ml/min per 1.73 m2 (hazard ratio 0.83 (0.77–0.91)) and a significantly increased risk at eGFR below 45 ml/min per 1.73 m2 (hazard ratio at 15 ml/min per 1.73 m2 1.93 (1.46–2.56)). In contrast, the relationship of albumin-to-creatinine ratio to the relative risk of all-cause mortality and cardiovascular mortality was monotonic with log hazard ratios increasing linearly with increasing log albumin-to-creatinine ratio, without threshold effects (Figure 1, right panels). As compared with an albumin-to-creatinine ratio of 5 mg/g, hazard ratios for all-cause mortality at albumin-to-creatinine ratios of 10, 30, and 300 mg/g were 1.08 (1.01–1.16), 1.38 (1.23–1.56), and 2.16 (1.99–2.35), respectively, and for cardiovascular mortality 1.13 (1.07–1.20), 1.55 (1.30–1.86), and 2.59 (1.95–3.44), respectively. The interaction between eGFR and albuminuria was significant for all-cause mortality in only 4 of 10 cohorts, and for cardiovascular mortality in only 1 of 7 cohorts (Supplementary Table S6 online). Significant interaction between eGFR and age was found in 3 of 10 cohorts for all-cause mortality, and in 2 out of 7 cohorts for cardiovascular mortality (Supplementary Table S6 online). Table 2 shows unadjusted incidence rates of all-cause and cardiovascular mortality for cohorts with albumin-to-creatinine ratio data. Pooled hazard ratios for all-cause mortality and cardiovascular mortality in the 28 categories of eGFR and albuminuria show that a higher albumin-to-creatinine ratio is associated with a higher risk across all levels of eGFR and a lower eGFR is associated with a higher risk across all levels of albumin-to-creatinine ratio, indicating multiplicative independence for all-cause mortality and cardiovascular mortality (Table 3). At severely reduced eGFR values (15–29 ml/min per 1.73 m2), the risk associated with higher albuminuria was slightly attenuated. The hazard ratios for all-cause mortality were significantly increased in all eGFR categories with an albumin-to-creatinine ratio ≥30 mg/g, whereas the hazard ratios for lower eGFR at an albumin-to-creatinine ratio <30 mg/g were significantly increased only for eGFR categories <45 ml/min per 1.73 m2. In contrast, the hazard ratios for cardiovascular mortality were significantly increased in all albumin-to-creatinine ratio ≥10 mg/g categories, even when eGFR was >60 ml/min per 1.73 m2 and for all eGFR <60 ml/min per 1.73 m2 categories, even when the albumin-to-creatinine ratio was <10 mg/g.Table 2Unadjusted incidence rates (per 1000 patient-years) for ACM and cardiovascular mortality in the high-risk cohorts with ACR ratio data Open table in a new tab Table 3Pooled adjusted hazard ratios (95% CI) for ACM and cardiovascular mortality in the high-risk cohorts with ACR dataView Large Image Figure ViewerDownload (PPT) Open table in a new tab For cohorts with dipstick data, unadjusted incidence rates for all-cause and cardiovascular mortality are shown in Table 4, and pooled hazard ratios for these end points in Table 5. This latter table shows that pooled hazard ratios for all-cause mortality were similarly increased for a higher dipstick category across all eGFR levels and for a lower eGFR across all dipstick categories. However, because of the U-shaped relationship, these hazard ratios were significantly higher than the reference group only at an eGFR level of 30–45 ml/min per 1.73 m2 and below, and hazard ratios were not significantly increased in the eGFR 45–59 ml/min per 1.73 m2 category with a negative or trace positive dipstick test.Table 4Unadjusted incidence rates (per 1000 patient-years) for ACM and cardiovascular mortality in the high-risk cohorts with dipstick data Open table in a new tab Table 5Pooled adjusted hazard ratios (95% CI) for ACM and cardiovascular mortality in high-risk cohorts with dipstick dataView Large Image Figure ViewerDownload (PPT) Open table in a new tab Figure 2 shows the continuous analyses (allowing interaction) of the hazard ratios of eGFR and albuminuria for all-cause mortality and cardiovascular mortality of cohorts with albumin-to-creatinine ratio and dipstick data. These figures suggest that eGFR and albuminuria are independently associated with all-cause mortality and cardiovascular mortality, with a tendency towards a weaker association between albumin-to-creatinine ratio and these outcomes at severely reduced eGFR values. The two cohorts with dipstick data that reported on cardiovascular mortality included very few persons with lower eGFR. Interpretation of these data is therefore difficult. The overall incidence rates for all-cause mortality and cardiovascular mortality were four- and twofold higher, respectively, in the subgroup of subjects with age ≥65 years compared with the subgroup with age <65 years (Supplementary Tables S7 and S8 online). For albumin-to-creatinine ratio cohorts, pooled hazard ratios for all-cause mortality and cardiovascular mortality of the 28 categories of eGFR and albumin-to-creatinine ratio according to age group are shown in Tables 6 and 7, respectively. For higher albumin-to-creatinine ratios, younger and older subjects showed a similar pattern for all-cause mortality and cardiovascular mortality. However, the patterns for eGFR were less steep among older subjects when compared with younger subjects. For all-cause mortality in dipstick cohorts, these patterns were generally similar to those in albumin-to-creatinine ratio cohorts (Table 8). Only one of the two dipstick cohorts that reported on cardiovascular mortality included subjects ≥65 years of age. For this reason, no data are shown for risk for cardiovascular mortality according to age group in dipstick cohorts.Table 6Pooled adjusted hazard ratios (95% CI) for ACM in the high-risk cohorts with ACR data, by age group Open table in a new tab Table 7Pooled adjusted hazard ratios (95% CI) for cardiovascular mortality in the high-risk cohorts with ACR data, by age groupView Large Image Figure ViewerDownload (PPT) Open table in a new tab Table 8Pooled adjusted hazard ratios (95% CI) for ACM in the cohorts with dipstick data, by age groupView Large Image Figure ViewerDownload (PPT) Open table in a new tab For the categorical analyses, we found modest but statistically significant heterogeneity in pooled adjusted hazard ratios within some eGFR × albuminuria categories. For all-cause mortality, heterogeneity was significant in 9 of 28 categories in the albumin-to-creatinine ratio cohorts, and in 6 of 28 categories in the dipstick cohorts (Supplementary Table S9 online). For cardiovascular mortality, heterogeneity was significant in 2 of 28 categories in the albumin-to-creatinine ratio cohorts (Supplementary Table S10 online). Qualitatively, however, the direction of the associations was the same in all cohorts—increased risk with lower eGFR categories and with higher albuminuria categories. Significant heterogeneity was, in nearly all cases, limited to the lowest eGFR and the highest albuminuria categories. Significant heterogeneity for all-cause mortality and cardiovascular mortality was observed in only very few of the eGFR × albuminuria categories of most clinical interest (eGFR of 45–60 ml/min per 1.73 m2 and albumin-to-creatinine ratio 30–300 mg/g or dipstick 1+). For the continuous analyses, forest plots are shown for the eGFR and albumin-to-creatinine ratio knots of most clinical interest (eGFR 45 ml/min per 1.73 m2 and albumin-to-creatinine ratio 30 mg/g) (Supplementary Figure S2 online). For all-cause mortality, significant heterogeneity was found, whereas for cardiovascular mortality, this was not observed. Meta-regression analysis was conducted for all-cause mortality and cardiovascular mortality at these knots to explore sources of heterogeneity, taking into account all variables listed in Table 1. Only two (near) significant associations were found: duration of follow-up and baseline eGFR was negatively associated with relative risk for all-cause mortality at an albumin-to-creatinine ratio 30 mg/g when compared with 5 mg/g (Supplementary Figure S3 online). However, both meta-regressions are, for a large part, explained by 1 outlier (Pima Indian study), characterized by a high baseline eGFR (144 ml/min per 1.73 m2). Without this outlier, no significant associations were noted (P=0.33 and P=0.20, respectively). In this collaborative meta-analysis of 10 high-risk cohorts, including 267,275 subjects, we found that a lower eGFR and a higher albuminuria were associated with a higher risk for all-cause mortality, independent of each other and independent of traditional cardiovascular disease risk factors. A similar association of eGFR and albuminuria was found with the risk for cardiovascular mortality. The risk for all-cause mortality and cardiovascular mortality based on eGFR and albuminuria has been reported in a limited number of high-risk cohorts.15.Ninomiya T. Perkovic V. de Galan B.E. et al.for the ADVANCE Collaborative Group. Albuminuria and kidney function independently predict cardiovascular and renal outcomes in diabetes.J Am Soc Nephrol. 2009; 20: 1813-1821Crossref PubMed Scopus (620) Google Scholar, 16.Hemmelgarn B.R. Manns B.J. Lloyd A. et al.Relation between kidney function, proteinuria, and adverse outcomes.JAMA. 2010; 303: 423-429Crossref PubMed Scopus (752) Google Scholar, 17.Tonelli M. Jose P. Curhan G. et al.Cholesterol and Recurrent Events (CARE) Trial Investigators. Proteinuria, impaired kidney function, and adverse outcomes in people with coronary disease: analysis of a previously conducted randomized trial.BMJ. 2006; 332: 1426-1432Crossref PubMed Scopus (168) Google Scholar The current meta-analysis confirms these studies and extends the generalizability of these data to other populations worldwide. Furthermore, our collaborative meta-analysis includes 16,702 all-cause mortality and 5656 cardiovascular mortality events, substantially more than the number of events in reports of individual studies, allowing more precise evaluation of the independent and joint associations of these measures with these outcomes. We observed an exponential increase in risk for both all-cause mortality and cardiovascular mortality risk at low eGFR. In the cohorts with albumin-to-creatinine ratio data, increased hazard ratios became statistically significant around eGFR 60 ml/min per 1.73 m2, and were two- and threefold higher at eGFR 45 and 15 ml/min per 1.73 m2, respectively, compared with optimal eGFR, independent of albuminuria and potential confounders. In the cohorts with dipstick data, we observed a U-shaped relationship, and the increased hazard ratio with lower eGFR became significant only below eGFR 45 ml/min per 1.73 m2. We also observed the U-shaped relationship with all-cause mortality in the general population cohorts,11.The Chronic Kidney Disease Prognosis Consortium Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality: a collaborative meta-analysis of general and high risk population cohorts.Lancet. 2010; 375: 2073-2081Abstract Full Text Full Text PDF PubMed Scopus (2460) Google Scholar and it has been observed by others.18.Coresh J. CKD prognosis: beyond the traditional outcomes.Am J Kidney Dis. 2009; 54: 1-3Abstract Full Text Full Text PDF PubMed Scopus (8) Google Scholar This U-shaped relationship can make it more difficult to interpreter the eGFR threshold for increased risk. Of note, the Modification of Diet in Renal Disease Study equation is known to underestimate measured GFR at the range of GFR ≥60 ml/min per 1.73 m2 in healthy individuals, and to have more variability among racial groups, such as Blacks and Pima Indians.19.Stevens L.A. Coresh J. Feldman H.I. et al.Evaluation of the modification of diet in renal disease study equation in a large diverse population.J Am Soc Nephrol. 2007; 18: 2749-2757Crossref PubMed Scopus (440) Google Scholar Furthermore, the Modification of Diet in Renal Disease equation overestimates measured GFR in individuals with reduced muscle mass because of ill health, the latter potentially contributing to the U-shaped association of GFR with mortality. The association of albuminuria with mortality was linear on the log-log scale, with a 1.5- and 2.5-fold higher risk at albumin-to-creatinine ratio 30 and 300 mg/g (corresponding to thresholds for microalbuminuria and macroalbuminuria), respectively, compared with an optimal albumin-to-creatinine ratio level (5 mg/g), independent of eGFR and conventional cardiovascular disease risk factors. Of note, the risk for cardiovascular mortality was statistically significant at an albumin-to-creatinine ratio of 10 mg/g compared with 5 mg/g. These findings are in agreement with previous reports that the association of albuminuria with all-cause mortality and cardiovascular mortality appears continuous with increased risk at levels below 30 mg/g.15.Ninomiya T. Perkovic V. de Galan B.E. et al.for the ADVANCE Collaborative Group. Albuminuria and kidney function independently predict cardiovascular and renal outcomes in diabetes.J Am Soc Nephrol. 2009; 20: 1813-1821Crossref PubMed Scopus (620) Google Scholar, 20.Hillege H.L. Fidler V. Diercks G.F. et al.for the PREVEND Study Group. Urinary albumin excretion predicts cardiovascular and noncardiovascular mortality in general population.Circulation. 2002; 106: 1777-1782Crossref PubMed Scopus (1287) Google Scholar, 21.Klausen K. Borch-Johnsen K. Feldt-Rasmussen B. et al.Very low levels of microalbuminuria are associated with increased risk of coronary heart disease and death independently of renal function, hypertension, and diabetes.Circulation. 2004; 110: 32-35Crossref PubMed Scopus (551) Google Scholar, 22.Gerstein H.C. Mann J.F. Yi Q. et al.for the HOPE Study Investigators. Albuminuria and risk of cardiovascular events, death, and heart failure in diabetic and nondiabetic individuals.JAMA. 2001; 286: 421-426Crossref PubMed Scopus (1925) Google Scholar Our findings of an increased relative risk of lower eGFR and of higher albuminuria were, in general, comparable for cohort studies with data on dipstick and cohort studies with data on albumin-to-creatinine ratio. These findings suggest that measurement of dipstick proteinuria is useful for risk stratification, despite being a less precise measure of albuminuria. The statistical code that was sent to the participating cohorts rendered output that did not permit computation of a meta-analytic result for interactions. However, the general pattern of a graded increase in risk with lower eGFR and higher albuminuria was present for subjects younger as well as older than 65 years of ag