estimation of glomerular filtration rate in diabetic subject

7/31/2019 Estimation of Glomerular Filtration Rate in Diabetic Subject

1/6

Estimation of Glomerular Filtration Ratein Diabetic SubjectsCockcroft formula or Modification of Diet in Renal Disease studyequation?

VINCENT RIGALLEAU, MD, PHD1

CATHERINE LASSEUR, MD2

CAROLINE PERLEMOINE, MD1

NICOLE BARTHE, PD3

C. RAFFAITIN, MD1

CHUNG LIU, MD1

PHILLIPE CHAUVEAU, MD2

LAURENCE BAILLET-BLANCO, MD1

MARIE-CHRISTINE BEAUVIEUX, PD1

C. COMBE, MD, PHD2

HENRI GIN, MD, PHD1

OBJECTIVE The Cockcroft-Gault formula is recommended for the evaluation of renal

function in diabetic patients. The more recent Modification of Diet in Renal Disease (MDRD)study equation seems more accurate, butit hasnot been validated in diabetic patients.This studycompares the two methods.

RESEARCH DESIGN AND METHODS In 160 diabetic patients, we compared theCockcroft-Gault formula and MDRD equation estimations to glomerular filtration rates (GFRs)measured by an isotopic method (51Cr-EDTA) by correlation studies and a Bland-Altman pro-cedure. Their accuracy for the diagnosis of moderately (GFR60 ml min1 1.73 m2) orseverely (GFR30ml min1 1.73m2) impaired renal function were compared with receiveroperating characteristic (ROC) curves.

RESULTS Both the Cockcroft-Gault formula (r 0.74; P 0.0001) and MDRD equation(r 0.81; P 0.0001) were well correlated with isotopic GFR. The Bland-Altman procedurerevealed a bias for the MDRD equation, which was not the case for the Cockcroft-Gault formula.Analysis of ROC curves showed that the MDRD equation had a better maximal accuracy for thediagnosisof moderate (areas under the curve [AUCs] 0.868 for the Cockcroft-Gault formula and0.927 forthe MDRD equation;P0.012)and severe renal failure (AUC 0.883 forthe Cockcroft-Gault formula and 0.962 for the MDRD equation; P 0.0001). In the 87 patients with renalinsufficiency, the MDRD equation estimation was better correlated with isotopic GFR (Cock-croft-Gault formula r 0.57; the MDRD equation r 0.78; P 0.01), and it was not biased asevaluated by the Bland-Altman procedure.

CONCLUSIONS Although both equations have imperfections, the MDRD equation ismore accurate for the diagnosis and stratification of renal failure in diabetic patients.

Diabetes Care 28:838843, 2005

Diabetic nephropathy affects 2540%of diabetic patients (1), and di-abetes is the leading cause of end-

stage renal disease (ESRD) in developedcountries (2). Mainly because of the highprevalence and increased life expectancyof type 2 diabetic patients (3), the propor-tion of patients with both diabetes andESRD is dramatically growing in devel-oped countries (4). Survival rates are low

in such patients because of high cardio-vascular risk (5), and medical costs arehigh (6).

The evaluation of renal function istherefore of critical importance in diabeticsubjects. Glomerular filtration rate (GFR)is the best measure of overall kidney func-tion in health and disease (7). Serum cre-atinine concentration is widely used as anindirect marker of GFR, but it is influ-enced by muscle mass and diet (8). GFRcan be directly measured by infusion ofexternal substances such as inulin or51Cr-EDTA (9), but these methods are ex-pensive and time consuming. The use ofprediction equations to estimate GFRfrom serum creatinine and other variables(age, sex, race, and body size) is thereforerecommended by the National KidneyFoundation for the diagnosis and stratifi-cation of chronic kidney diseases (10).

According to these guidelines, renal func-tion is moderately decreased if GFR is60 ml min1 1.73 m2 and severelydecreased if GFR is30ml min1 1.73m2.

The proposed equations are the

Cockcroft-Gault formula (11), as recom-mended by the American Diabetes Asso-ciation (12), and the Modification of Dietin Renal Disease (MDRD) study equation(13). The more recent MDRD equationseems more accurate, but it has not beenvalidated in diabetic kidney disease (10).Its superiority over the Cockcroft-Gaultformula has been mentioned in some(14), but not all (15,16), recent reports.

We compared Cockcroft-Gault for-mula and MDRD equation estimates ofGFR with 51Cr-EDTA measurement in

From the 1Department of Nutrition and Diabetes, Hopital Haut-Leveque, Pessac, France; the 2Departmentof Nephrology, University Victor Segalen Bordeaux II, Bordeaux, France; and the 3Department of NuclearMedicine, University Victor Segalen Bordeaux II, Bordeaux, France.

Address correspondence and reprint requests to Vincent Rigalleau, Nutrition-Diabetologie, Hopital Haut-Leveque, Avenue de Magellan, 33600 Pessac, France. E-mail: [email protected].

Received for publication 12 September 2004 and accepted in revised form 20 December 2004.Abbreviations:AUC, area under the curve; ESRD, end-stage renal disease; GFR, glomerular filtration rate;

MDRD, Modification of Diet in Renal Disease; ROC, receiver operating characteristic.A table elsewhere in this issue shows conventional and Systeme International (SI) units and conversion

factors for many substances. 2005 by the American Diabetes Association.The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby

marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

P a t h o p h y s i o l o g y / C o m p l i c a t i o n sO R I G I N A L A R T I C L E

838 DIABETES CARE, VOLUME 28, NUMBER 4, APRIL 2005


2/6

160 diabetic subjects with a wide range ofGFRs (8164 ml min1 1.73 m2).Westudied the correlation between both es-timations and isotopic measurement ofGFR and performed a Bland-Altman pro-cedure (17). Their sensitivity and speci-ficity for the diagnosis of moderately orseverely impaired renal function were as-sessed from receiver operating character-istic curves (ROC). Correlation studiesand Bland-Altman procedures were alsoperformed on the renal insufficient groupin relation to isotopic GFR (n 87).

RESEARCH DESIGN AND

METHODS The study group con-sisted of 160 diabetic patients attendingour clinical unit. Both sexes (91 men and69 women) and types of diabetes (50 type

1 and 110 type 2) were represented.Mean SD HbA

1cwas 8.6 1.7%. A

wide range of ages (19 83 years; 62.213.7 years), BMIs (15.648.9 kg/m2;27.5 4.6 kg/m2), and serum creatininelevels (54371 mol/l; 136.0 69.1mol/l) were represented. Mean protein-uria was 523 260 mg/24 h. Subjectswith nephrotic proteinuria (3g/24h)orclinical edema were excluded. No subjectwas treated by dialysis at the time of thestudy.

Serum creatinine was determined on

a multiparameter analyzer (Olympus AU640; Olympus Optical, Tokyo, Japan) us-ing the Jaffe method with bichromaticmeasurements according to the manufac-turers specifications. Clearance of the ra-dionuclide marker was measured afterintravenous injection of51Cr-EDTA (CisIndustries, Gif/Yvette, France). All pa-tients were studied at 9:00 A.M. after alight breakfast. After a single bolus of 100Ci (3.7 MBq) of 51Cr-EDTA, four ve-nous blood samples were drawn at 75,105, 135, and 165 min and urinary sam-

ples were collected at 90, 120, 150, and180 min as previously described (18).The 51Cr-EDTA radioactivity was mea-sured on a counter (COBRA 2, model05003; Packard Instruments, Meriden,CT).

Estimations of renal functionA single creatinine determination wasperformed the day before the isotopicmeasurement of GFR to calculate theCockcroft-Gault formula as follows.

140 age [years]) body weight (kg) K

serum creatinine (mol/l)

(1)

where K is a constant of 1.23 for men and1.04 for women (11).

Before comparison, Cockcroft-Gaultformula results were adjusted to bodysurface area using Dubois formula (19).

To calculate the MDRD equation, weused the following abbreviated equation(13).

186 (serum creatinine [mg/dl])1.154

([years])0.203 (0.742 if female)

(1.210 if African American)

Statistical analysisResults of the Cockcroft-Gault formula

and the MDRD equation were comparedwith isotopic GFR by correlation, paired ttests, and Bland-Altman procedures.These calculations were performed withSPSS software, version 10.0. The sensitiv-ity and specificity of both formulas wereassessed from nonparametric ROC curvesgenerated by plotting sensitivity versus1 specificity, giving the ideal test a sen-sitivity 1 and specificity 1. Areas un-der the curve (AUCs) were calculated andcompared according to the procedure ofHanley and McNeil (20). AUC is com-monly 0.5 with values ranging from 1(ideal perfect separation of the tested val-ues) to 0.5 (no apparent distribution dif-ference between the tested groups). Theseanalyses were performed using Medcalcsoftware. Results are presented asmeans SD. P 0.05 was consideredsignificant.

RESULTS Mean isotopic GFR was60.9 36.3 ml min1 1.73 m2. Themean Cockcroft-Gault formula overesti-mated GFR (65.6 37.5 ml min1 1.73 m2; P 0.05 vs. isotopic GFR) and

the mean MDRD equation underesti-mated GFR (54.7 25.1 ml min1 1.73 m2; P 0.001 vs. isotopic GFR).

As shown in Fig. 1A, both estimationswere well correlated to isotopic GFR witha slight advantage for the MDRD equation(Cockcroft-Gault formula r 0.74, P 0.0001; MDRD equation r 0.81, P 0.0001; P 0.12 between rvalues). TheBland-Altman procedure (Fig. 1B) re-vealed a bias for the MDRD equation astheestimation minus GFR (mean6.1ml min1 1.73 m2, 2 SDs 43.0) was neg-

atively correlated to the mean (r 0.54, P 0.001), which was not thecase for the Cockcroft-Gault formula(mean4.8 ml min1 1.73 m2,2SDs52.4, r 0.04, P 0.64).

For the 50 type 1 diabetic subjects(BMI 24.6 2.9kg/m2), both estimations

did not differ from isotopic GFR (isotopic65.5 34.2 ml min1 1.73 m2; Cock-croft-Gault formula 66.6 35.4 ml min1 1.73 m2; MDRD equation62.4 29.7 ml min1 1.73 m2; NS).For the 110 type 2 diabetic subjects (BMI28.9 4.8 kg/m2; P 0.0001 vs. type 1),isotopic GFR was 58.7 37.2 ml min1

1.73 m2. The Cockcroft-Gault formulagave an overestimated GFR (65.2 38.5ml min1 1.73 m2; P 0.01 vs. iso-topic) and the MDRD equation gave anunderestimated GFR (51.2 22.0 ml

min1

1.73 m

2

; P 0.001 vs. isoto-pic). In both subgroups, correlation coef-ficients between estimated and measuredGFR were not significantly, but were con-sistently, higher for the MDRD equation(type 1: r 0.72 for the Cockcroft-Gaultformula and 0.83 for the MDRD equation;type 2: r 0.76 for the Cockcroft-Gaultformula and 0.83 for the MDRD equa-tion).

The ROC curve analysis (Fig. 2)showed that the maximum diagnostic ac-curacy of the Cockcroft-Gault formula forthe diagnosis of moderate renal failure(GFR60 ml min1 1.73 m2) waslower than the MDRD equation (Cock-croft-Gault formula AUC 0.868, cutofflimit 56.5; MDRD equation AUC 0.927,cutoff limit 54.7; P 0.05). This wasmainly due to a better sensitivity of theMDRD equation estimation (Cockcroft-Gault formula sensitivity 77.9% and spec-ificity 81.1%; MDRD equation sensitivity91.9% and specificity: 78.4%). For the di-agnosis of severe renal failure (GFR30ml min1 1.73 m2) (Fig. 3), the max-imum diagnostic accuracy of the Cock-

croft-Gault formula waslower than that ofthe MDRD equation (Cockcroft-Gaultformula AUC 0.883, cutoff limit 43.9;MDRD equation AUC 0.962, cutoff limit42.4; P 0.0001) because of improvedsensitivity and specificity (Cockcroft-Gault formula sensitivity 78.9% and spec-ificity 84.4%; MDRD equation sensitivity94.7% and specificity 90.2%).

In the 87 renal-insufficient patients(GFR60 ml min1 1.73 m2; meanisotopic GFR 33.7 14.7 ml min1 1.73 m2), both the Cockcroft-Gault for-

Rigalleau and Associates

DIABETES CARE, VOLUME 28, NUMBER 4, APRIL 2005 839


3/6

mula and the MDRD equation overesti-mated GFR (Cockcroft-Gault formula45.3 21.0 ml min1 1.73 m2;MDRD equation 38.4 14.0 ml min1

1.73 m2; both P 0.0001 vs. isotopicGFR), but the overestimation was morepronounced with the Cockcroft-Gaultformula (P 0.0001 vs. the MDRD equa-tion). As shown inFig. 4A, the correlationwith isotopic GFR was lower for theCockcroft-Gault formula (Cockcroft-Gault formula r 0.57, P 0.001;MDRD equation r 0.78, P 0.0001;P 0.01 between rvalues). As shown inFig. 4B, the Cockcroft-Gault formulaoverestimated high values of GFR accord-ing to the Bland-Altman procedure (mean11.6 ml min1 1.73 m2, 2SD 37.1,r 0.38, P 0.001); this wasnot thecasefor the MDRD equation (mean 4.7 ml

min1

1.73 m

2

, 2SD 20.6, r 0.07,P 0.49).

CONCLUSIONS The Cockcroft-Gault formula is a simple, widely used,and recommended means to assess renalfunction. The estimation by the Cock-croft-Gault formula is well correlated (r0.750.93) with GFR as determined byinfusion of external substances such as99Tc-diethylenetriaminepentaacetate(21,22), iothamalate (13,2326), inulin(27), iohexol (28), and 51Cr-EDTA (29).

Studies in diabetic patients (24,3032)have examined smaller numbers of pa-tients (n 49136) and reported slightlylower correlation coefficients (r 0.690.88). A correlation coefficient of 0.74and a slight overestimation of GFR (24),as we also found, were therefore not un-expected. The normalization of the Cock-croft-Gault formula to body surface areaas carried out is known to improve its di-agnostic performance (33).

Previous studies found the mainproblem with the Cockcroft-Gault for-

mula to be overestimation when GFR val-ues are low (21,22), which we also foundto be true, particularly in the 87 diabeticpatients with renal insufficiency. Our re-sults show that this overestimation altersthe sensitivity of the Cockcroft-Gault for-mula for the diagnosis of moderate andsevere renal failure. Late diagnosis of se-vere renal failure can retard the referral tothe nephrologist or the indication for di-alysis or transplantation, which worsensthe prognosis (34). The National KidneyFoundation recommends that patients

Figure 1A: Estimated GFR as a function of its isotopic measurement (milliliters per minute per 1.73 m2) in 160 diabetic subjects. B: Bland-Altmanplots of differences between estimated GFR and measured with 51Cr-EDTA as a function of average GFR by both methods in 160 diabetic subjects.E, Cockcroft-Gault formula; F, MDRD equation.

Figure 2ROC curves comparing AUCs ofthe Cockcroft-Gault formula and the MDRDequation for the diagnosis of moderate renalfailure (GFR 60 ml min1 1.73 m2).

Evaluation of renal function in diabetes



4/6


5/6

formula. However, an important practicalutility of a GFR predictive formula is todiagnose and stratify chronic renal failurein patients with renal diseases. Our workshows better precision and improved di-agnostic accuracy for the MDRD equa-tion, particularly in diabetic subjects with

renal insufficiency. This advantage maybe offset by the requirement of a scientificcalculator because the MDRD equationcalculation includes negative logarithms,whereas the Cockcroft-Gault formula canbe calculated using a simple calculator orby mental arithmetic. As suggested by theNational Kidney Foundation recommen-dations (10), the calculation of the esti-mated GFR can be performed by clinicallaboratories; according to French recom-mendations (33), laboratories give an es-timation of GFR together with the result

of serum creatinine measurement usingthe Cockcroft-Gault formula. If other re-ports confirm the practical interest of theMDRD equation, it will be possible to re-quest a MDRD equation calculation fromthe laboratory.

Although the error is halved whencompared with the Cockcroft-Gault for-mula, low GFR was still overestimatedwith the MDRD equation. The relation-ship between serum creatinine and GFR isnot simple in severe renal insufficiency.The clearance of creatinine does not de-pend only on GFR (38) as tubular excre-tion of creatinine occurs. This tubularexcretion varies with the degree of renalfailure (39), thus limiting the possibilityof using 24-h creatinine clearance, whichis not recommended by the National Kid-ney Foundation as a substitute for predic-tion equations (10). The serum level ofcreatinine not only depends on its clear-ance; lower creatinine production due todecreased lean body mass (40) may re-duce creatinine production in some renalinsufficient patients as renal function de-clines. A lower creatinine generation has

indeed been mentioned in hemodialyzeddiabetic subjects (35). Further improvedformulas are therefore warranted and theuse of reference methods with infusion ofexternal substances can still be useful toevaluate renal function in some renal-insufficient diabetic patients.

In summary, both equations have im-perfections. Overestimation at low GFRlevels and the influence of weight reducethe sensibility and the accuracy of theCockcroft-Gault formula. The MDRDequation is more difficult to calculate in

clinical practice and underestimates GFRat high levels, but it has better accuracy indiagnosing and stratifying chronic renalfailure in diabetic patients, which is an im-portant advantage for a prediction formula.

References1. Parving HH: Renoprotection in diabetes:

genetic and non-genetic risk factors andtreatment. Diabetologia 41:745759, 1998

2. National Technical Information Service:UnitedStates Renal Data SystemAnnual Re-port. Springfield, VA, U.S. Dept. of Healthand Human Services, 1999

3. Ritz E, Orth SR: Nephropathy in patientswith type 2 diabetes mellitus.NEnglJMed341:11271133, 1999

4. RitzE, Rychlik I, Wahl P, Michael C:End-stage renal failure in type 2 diabetes: amedical catastrophe of worldwide dimen-

sion. Am J Kidney Dis 34:795808, 19995. Chantrel F, Enache I, Bouiller M, Kolb I,Kunz K, Petitjean P, Moulin B, Hanne-douche T: Abysmal prognosis of patientswith type 2 diabetes entering dialysis.Nephrol Dial Transplant 14:129136,1999

6. Smith DG, Harlan LC, Hawthorne VM:The charges forESRD treatment of diabet-ics. J Clin Epidemiol 42:111118, 1989

7. Smith HW: Comparative physiology ofthe kidney. In The Kidney: Structure andFunction in Health and Disease. Smith HW,Ed. New York, Oxford University Press,1951, p. 52505274

8. Levey AS, Perrone RD, Madias NE: Serumcreatinine and renal function. Annu RevMed 39:465490, 1988

9. Heath DA, Knapp MS, Walker WH: Com-parison between inulin and 51Cr-labellededetic acid for the measurement of glo-merular filtration rate. Lancet 2:11101112, 1968

10. Levey AS, Coresh J, Balk E, Kausz AT,Levin A, Steffes MW, Hogg RJ, PerroneRD, Lau J, Eknoyan G: National KidneyFoundation practice guidelines forchronic kidney disease: evaluation, classi-fication and stratification. Ann Intern Med139:137147, 2003

11. Cockcroft DW, Gault HM: Prediction ofcreatinine clearance from serum creati-nine. Nephron 16:3141, 1976

12. American Diabetes Association: Stan-dards of medical care for patients with di-abetes mellitus. Diabetes Care 24:S33S43, 2001

13. Levey AS, Bosch JP, Lewis JB, Greene T,Rogers N, Roth D: A more accuratemethod to estimate glomerular filtrationrate from serum creatinine: a new predic-tion equation: Modification of Diet in Re-nal Disease Study Group. Ann Intern Med130:461470, 1999

14. Hallan S, Asberg A, Lindberg M, JohnsenH: Validation of the Modification of Dietin Renal Disease formula for estimatingGFR with special emphasis on calibrationof the serum creatinine assay.Am J KidneyDis 44:8493, 2004

15. Vervoort G, Willems HL, Wetzels JF: As-

sessment of glomerular filtration rate inhealthy subjects and normoalbuminuricdiabetic patients: validity of a new(MDRD) prediction equation. NephrolDial Transplant 17:19091913, 2002

16. Lamb EJ, Webb MC, Simpson DE, Coak-ley AJ, Newman DJ, ORiordan SE: Esti-mation of glomerular filtration rate inolder patients with chronic renal insuffi-ciency: is the Modification of Diet in RenalDisease formula an improvement? J AmGeriatr Soc 51:10121017, 2003

17. Bland JM, Altman DG: Statistical methodsfor assessing agreement between twomethods of clinical measurement. Lancet

2:307310, 198618. Combe C, Deforges-Lasseur C, Caix J,Pommereau A, Marot D, Aparicio M:Compliance and effects of nutritionaltreatment on progression and metabolicdisorders of chronic renal failure. NephrolDial Transplant 8:412418, 1993

19. DuBois D, DuBois EF: A formula to esti-mate the approximate surface area ifheight and weight are known. Ann InternMed 17:863871, 1916

20. Hanley JA, McNeil BJ: The meaning anduse of the area under a receiver operatingcharacteristic (ROC) curve. Radiology143:2936, 1982

21. Waller DG, Flemming JS, Ramsey B, GrayJ: The accuracy of creatinine clearancewith and without urine collection as ameasure of glomerularfiltration rate. Post-grad Med J67:4246, 1991

22. Gault MH, Longerich LL, Harnett JD,Wesolowski C: Predicting glomerularfunction from adjusted serum creatinine.Nephron 62:249256, 1992

23. Rolin HA, Hall PM, Wei R: Inaccuracy ofestimated creatinine clearance for predic-tion of iothalamate glomerular filtrationrate. Am J Kidney Dis 4:4854, 1984

24. Lemann J, Bidani AK, Bain RP, Lewis EJ,Rohde RD: Use of the serum creatinine to

estimate glomerular filtration rate inhealth and early diabetic nephropathy.Am J Kidney Dis 16:236243, 1990

25. Toto RD, Kirk KA, Coresh J, Jones C,Appel L, Wright J, Campese V, Olutade B,Agodoa L: Evaluation of serum creatininefor estimating glomerular filtration rate inAfrican Americans with hypertensive ne-phrosclerosis: results from the African-American study of Kidney Disease andHypertension (AASK) pilot study. J AmSoc Nephrol 8:279287, 1997

26. Lewis J, Agodoa L, Cheek D, Greene T,Middleton J, OConnor D, Ojo A, Phillips

Evaluation of renal function in diabetes



6/6

estimation of glomerular filtration rate in diabetic subject

Documents