Adding Measures of Body Composition to the CKD-EPI GFR Estimating Equation in Indigenous Australians: The eGFR Study

Abstract

The accuracy of the CKD-EPI equation to estimate GFR may be limited in ethnic groups not included in the CKD-EPI development population due to differences in creatinine generation (related to differences in muscle mass and/or diet). Aboriginal Australians traditionally had a “linear” body build (narrow across the shoulders and hips; relatively long limbs and short torso), which was associated with relatively less muscle and more fat for a given weight.1Rutishauser I.H. McKay H. Anthropometric status and body composition in aboriginal women of the Kimberley region.Med J Aust. 1986; 144: S8-S10PubMed Google Scholar, 2Piers L.S. Rowley K.G. Soares M.J. O'Dea K. Relation of adiposity and body fat distribution to body mass index in Australians of aboriginal and European ancestry.Eur J Clin Nutr. 2003; 57: 956-963Crossref PubMed Scopus (78) Google Scholar Although some studies have assessed the utility of incorporating measures of FFM to estimate GFR, limited data are available.3Taylor T.P. Wang W. Shrayyef M.Z. Cheek D. Hutchison F.N. Gadegbeku C.A. Glomerular filtration rate can be accurately predicted using lean mass measured by dual-energy x-ray absorptiometry.Nephrol Dial Transplant. 2006; 21: 84-87Crossref PubMed Scopus (29) Google Scholar, 4Macdonald J.H. Marcora S.M. Kumwenda M.J. et al.The relationship between estimated glomerular filtration rate, demographic and anthropometric variables is mediated by muscle mass in non-diabetic patients with chronic kidney disease.Nephrol Dial Transplant. 2006; 21: 3488-3494Crossref PubMed Scopus (32) Google Scholar, 5Macdonald J.H. Marcora S.M. Jibani M. et al.Bioelectrical impedance can be used to predict muscle mass and hence improve estimation of glomerular filtration rate in non-diabetic patients with chronic kidney disease.Nephrol Dial Transplant. 2006; 21: 3481-3487Crossref PubMed Scopus (43) Google Scholar A measure of FFM may be obtained by measuring body composition using the simple, portable method of bioelectrical impedance analysis (BIA), which we recently validated against dual-energy x-ray absorptiometry (DEXA).6Hughes J.T. Maple-Brown L. Piers L.S. Meerkin J. O'Dea K. Ward L.C. Development of a single frequency bioimpedance prediction equation for fat-free mass in an adult indigenous Australian population.Eur J Clin Nutr. 2015; 69: 28-33Crossref PubMed Scopus (9) Google Scholar We hypothesized that GFR might be estimated more accurately with an equation involving Scr and an estimate of FFM as this would adjust for variation in creatinine production related to variation in lean body mass (LBM). The aim of this study was to assess whether incorporating an FFM estimate or other anthropometric measurements into the prediction equation would be likely to significantly improve GFR estimation in Indigenous Australians. It is not this article’s goal to derive a new equation for routine clinical use in Indigenous Australians as the original CKD-EPI equation is suitable7Maple-Brown L.J. Hughes J.T. Lawton P.D. et al.Accurate assessment of kidney function in indigenous Australians: the Estimated GFR Study.Am J Kidney Dis. 2012; 60: 680-682Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar and the current study does not provide a validation set. The eGFR Study’s methods have been reported previously.7Maple-Brown L.J. Hughes J.T. Lawton P.D. et al.Accurate assessment of kidney function in indigenous Australians: the Estimated GFR Study.Am J Kidney Dis. 2012; 60: 680-682Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar, 8Maple-Brown L.J. Lawton P.D. Hughes J.T. et al.Study protocol—accurate assessment of kidney function in indigenous Australians: aims and methods of the eGFR study.BMC Public Health. 2010; 10: 80Crossref PubMed Scopus (30) Google Scholar Indigenous Australian participants 16 years or older were recruited across 5 predefined strata of health, diabetes status, and kidney function across 4 large geographic regions of Australia (Item S1). A total of 564 participants 18 years or older underwent measurement of GFR (4-hour elimination of plasma iohexol), Scr (by an enzymatic method), and weight; 483 of these had complete data for body composition variables, including BIA. Linear regression models of normalized mGFR were fitted, including Scr, sex, and age in the same functional form (but different coefficients) as the CKD-EPI equation, and various measures of body composition. Participant characteristics (Item S1) show differences in body composition and eGFR between sexes. Regression models estimated GFR on participants with complete data (Table 1). Adding weight led to a small improvement (3.6% reduction in RMSE); however, inclusion of body composition, height, or waist or hip circumference gave negligible further benefit. The resulting main model was derived using age, sex, Scr, and weight: eGFR (mL/min/1.73 m2) = 165 × minimum(Scr/k, 1)α × maximum(Scr/k, 1)-0.99 × 0.9927age × 0.98 [if female] × exp(−573/weight2), where k is 62 (females) or 80 (males) and α is −0.431 (females) or −0.533 (males).Table 1Results of Regression Models of mGFR (mL/min/1.73 m2)aLog transformed. for 483 Indigenous Participants With Complete DataLocal CKD-EPI ModelWeight ModelWeight & Height ModelWeight, Height, Waist, & Hip ModelWeight, Height & FFM ModelScr & FFM Only ModelScr (μmol/L)++++++Age (y)+++++−Female+++++−WeightaLog transformed. (kg)−++++−HeightaLog transformed. (m)−−+++−WaistaLog transformed. circum (cm)−−−+−−HipaLog transformed. circum (cm)−−−+−−FFMaLog transformed. (kg)−−−−++RMSE0.17210.16590.16560.16460.16500.1896ΔRMSEbRelative to CKD-EPI model.—3.6%3.8%4.4%4.1%−10.2%Model R20.82840.84090.84180.84440.84330.792Abbreviations and definitions: +, inclusion of the variable in the model; circum, circumference; CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration; FFM, fat-free mass; mGFR, measured glomerular filtration rate; RMSE, root mean square error; Scr, serum creatinine.a Log transformed.b Relative to CKD-EPI model. Open table in a new tab Abbreviations and definitions: +, inclusion of the variable in the model; circum, circumference; CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration; FFM, fat-free mass; mGFR, measured glomerular filtration rate; RMSE, root mean square error; Scr, serum creatinine. The influence of weight on GFR in the main model (Item S1) was greatest at low weights. Without weight in the model, differences in bias were seen between those with or without low (<72.5 kg) weight (Fig 1A). These differences disappeared when weight was included in the main model (Fig 1B). Previously, equations with weight have showed an improvement in those with BMI < 20 kg/m2.9Stevens L.A. Schmid C.H. Greene T. et al.Comparative performance of the CKD Epidemiology Collaboration (CKD-EPI) and the Modification of Diet in Renal Disease (MDRD) Study equations for estimating GFR levels above 60 mL/min/1.73 m2.Am J Kidney Dis. 2010; 56: 486-495Abstract Full Text Full Text PDF PubMed Scopus (473) Google Scholar In contrast, addition of LBM (estimated using weight, height, and sex) resulted in improvement in an obese male subgroup of the Lund-Malmö study.10Björk J. Bäck S.E. Sterner G. et al.Prediction of relative glomerular filtration rate in adults: new improved equations based on Swedish Caucasians and standardized plasma-creatinine assays.Scand J Clin Lab Invest. 2007; 67: 678-695Crossref PubMed Scopus (69) Google Scholar Our finding that FFM did not improve the model to estimate GFR is consistent with that of Rule et al11Rule A.D. Bailey K.R. Schwartz G.L. Khosla S. Lieske J.C. Melton III, L.J. For estimating creatinine clearance measuring muscle mass gives better results than those based on demographics.Kidney Int. 2009; 75: 1071-1078Crossref PubMed Scopus (93) Google Scholar (which measured urinary creatinine and muscle mass [by DEXA] in non-Hispanic white Americans). However, addition of weight and appendicular lean mass (using bioimpedance) improved bias and accuracy of eGFR (mL/min) in 75 nondiabetic CKD patients in Macdonald et al.5Macdonald J.H. Marcora S.M. Jibani M. et al.Bioelectrical impedance can be used to predict muscle mass and hence improve estimation of glomerular filtration rate in non-diabetic patients with chronic kidney disease.Nephrol Dial Transplant. 2006; 21: 3481-3487Crossref PubMed Scopus (43) Google Scholar Comparison of our study with that of Macdonald et al5Macdonald J.H. Marcora S.M. Jibani M. et al.Bioelectrical impedance can be used to predict muscle mass and hence improve estimation of glomerular filtration rate in non-diabetic patients with chronic kidney disease.Nephrol Dial Transplant. 2006; 21: 3481-3487Crossref PubMed Scopus (43) Google Scholar is limited by different body composition determination methods and participant characteristics (mean mGFRs, 103 and 45.7 mL/min/1.73 m2, respectively). Our participants were volunteers recruited across strata of health, diabetes status, and kidney function so we cannot comment on how representative they are of their communities. Also, GFR measurement by 4-hour iohexol clearance can have reduced accuracy in individuals with decreased GFR. Nevertheless, a key strength is the large number of participants with reference mGFRs across a range of health, diabetes, and kidney function. In conclusion, addition of weight to the established inputs of age, sex, and Scr slightly enhanced eGFR equation performance (particularly at low body weight). The scatter of formal GFR measurements compared with eGFR prediction equations remains large despite consideration of other body measurement variables. We recommend continued use of the CKD-EPI equation in routine clinical practice. Further work using cystatin C (including with Scr) may improve precision in this high-risk population, as has been reported in Asian populations and subgroups, including those of low BMI.12Teo B.W. Xu H. Wang D. et al.Estimating glomerular filtration rates by use of both cystatin C and standardized serum creatinine avoids ethnicity coefficients in Asian patients with chronic kidney disease.Clin Chem. 2012; 58: 450-457Crossref PubMed Scopus (28) Google Scholar, 13Fan L. Inker L.A. Rossert J. et al.Glomerular filtration rate estimation using cystatin C alone or combined with creatinine as a confirmatory test.Nephrol Dial Transplant. 2014; 29: 1195-1203Crossref PubMed Scopus (64) Google Scholar We thank the participants, staff, and investigators of the eGFR Study. The authors write on behalf of the study investigators, who include Mark Thomas, Alex Brown, Ashim Sinha, Robyn MacDermott, Kevin Warr, Sajiv Cherian, and William Majoni. Support: The eGFR Study was funded by the National Health and Medical Research Council (NHMRC; project grant 545202), with additional support from Kidney Health Australia, NHMRC#320860, the Colonial Foundation, Diabetes Australia Research Trust, and the Rebecca L. Cooper Foundation. LM-B is supported by an NHMRC Early Career Fellowship in Aboriginal and Torres Strait Islander Health Research (605837); JH, by NHMRC Scholarship 490348, Rio Tinto Aboriginal Fund, and the Centre of Clinical Research Excellence in Clinical Science of Diabetes, University of Melbourne; PL, by NHMRC Scholarship 1038529. AC holds NHMRC Principal Research Fellowship 1027204; WH, NHMRC Australia Fellowship 511081. We thank Roche Diagnostics for enzymatic creatinine reagents, Melbourne Pathology for technical support in enzymatic creatinine analysis, SeaSwift for in-kind support in transportation of the DEXA to Thursday Island, and MeasureUp for assisting in having their mobile bone densitometry vehicle present on Thursday Island. Funding bodies had no role in the study design; collection, analysis, or interpretation of data; or the writing or decision to submit the manuscript. Financial Disclosure: LCW consults for Impedimed Ltd, which had no involvement in the concept, design, or execution of this study or preparation of this manuscript. The remaining authors declare that they have no other relevant financial interests. Contributions: Research idea and study design: LMB, JH, LCW, LSP, PDL, GRDJ, AGE, WH, AC, KOD, RJM, GJ; data acquisition: LMB, JH; laboratory measurement of mGFR: AGE; data analysis/interpretation: LMB, MC; statistical analysis: MC; supervision or mentorship: LMB, KOD, GJ. Each author contributed important intellectual content during manuscript drafting or revision and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of the work are appropriately investigated and resolved. LMB takes responsibility that this study has been reported honestly, accurately, and transparently; that no important aspects of the study have been omitted, and that any discrepancies from the study as planned have been explained. Download .pdf (.16 MB) Help with pdf files Supplementary Item S1 (PDF)Methods, participant characteristics, and modeled effect of weight on GFR.