Sage Journals: Discover world-class research

Abstract

Background

Estimation of glomerular filtration rate (GFR) using plasma creatinine remains controversial, especially when GFR approaches normal values. The aim was to re-examine estimated GFR (eGFR) using dual-reference GFR measurements.

Methods

eGFR (simplified modified Modification of Diet in Renal Disease equation) was compared with GFR measured with iohexol for predicting GFR measured with ⁵¹Cr-ethylenediaminetetraacetic acid (EDTA). Dual six-sample GFR (20–240 min postinjection) was measured in 60 patients and 20 normal volunteers with ⁵¹Cr-EDTA (GFR_EDTA) and iohexol (GFR_iohexol) injected into separate arms and sampled contralaterally. This was repeated in the normal volunteers under fasting conditions (twice in nine). Percentage bias, imprecision (SD of bias) and disagreement (sign-less difference) between eGFR and GFR_EDTA were compared with those between GFR_iohexol and GFR_EDTA.

Results

Changes between fasting and postprandial eGFR correlated significantly with corresponding changes in GFR_iohexol and GFR_EDTA. eGFR predicted GFR_EDTA less precisely (SD 19.9%) than GFR_iohexol (10.5%; P < 0.01). Although eGFR showed a poorer correlation with GFR_EDTA when GFR_EDTA > 80 mL/min/1.73 m² compared with <80 mL/min/1.73 m², there was no significant difference with respect to imprecision or disagreement of >20 or 30%. However, eGFR was closer than GFR_iohexol to GFR_EDTA in a higher fraction of studies when GFR_EDTA > 80 mL/min/1.73 m² (28/60) than when it was <80 mL/min/1.73 m² (9/37; P < 0.05).

Conclusion

eGFR is inferior to GFR_iohexol for predicting GFR_EDTA. The disagreement between GFR_iohexol and GFR_EDTA illustrates the extent to which uncertainty in GFR_EDTA contributes to the performance of eGFR. eGFR performs no better at lower, compared with higher levels of GFR.

Introduction

There have been numerous evaluations of the Modification of Diet in Renal Disease (MDRD) study equations for the estimation of glomerular filtration rate (eGFR) from plasma creatinine concentration.¹ Most find eGFR of value, both in general populations² and in special groups such as the elderly,^3,4 diabetics^5,6 and patients with chronic kidney disease (CKD)^7–11 but there are several reporting no improvement over other creatinine-based methods or poor performance in transplant recipients,¹² patients with CKD,^13,14 including those with normal creatinine,¹⁵ diabetics^16–19 and the obese,³ or misclassification of patients with renal disease into the wrong CKD stage.²⁰ Moreover, variations in creatinine assay,^21,22 standardization of the assay^23,24 and a preceding meal^25,26 also affect the predictive power of eGFR. The only well-studied ethnic groups are Caucasians and Afro-Caribbeans, although there have been recent studies in Indians,²⁷ Japanese²⁸ and Chinese.^11,13 The MDRD equations are neither considered appropriate in children²⁹ nor when GFR exceeds 60 mL/min/1.73 m².^{10,16,18,27,30–33} or 90 mlLmin/1.73 m².^2,4

Evaluations of eGFR have been based on numerous reference indicators, including radiopharmaceuticals (⁵¹Cr-ethylenediaminetetraacetic acid [EDTA],^{5,14,17,18,21,22} ^{99 m}Tc-diethylenetriaminepentaacetic [DTPA]^{3,11,13,27,30,34} and ¹²⁵I-iothalamate^2,34) and stable compounds (iohexol¹⁵, inulin^4,8,9,16,29 and iothalamate^7,31) and on numerous measurements, including non-steady-state two-compartment or one-compartment plasma and urinary clearances and steady-state urinary clearance. Any particular reference technique, however, has its own experimental error and what is missing from the literature is a comparison between eGFR and a reference method that at the same time is also compared with a second, independent reference method. Such an inclusion would highlight the imprecision and inaccuracy of the gold standard methods and show to what extent differences between eGFR and either reference method are exclusively the result of issues with eGFR rather than the reference methods themselves.

The first aim of our study, therefore, was to evaluate eGFR initially by comparing its ability to predict multisample (two-compartment) GFR measured with ⁵¹Cr-EDTA (GFR_EDTA; the primary gold standard) with that of multisample GFR simultaneously and independently measured with iohexol (GFR_iohexol; the secondary gold standard), a filtration marker that has been validated against both ⁵¹Cr-EDTA^35–39 and inulin^40–43). The second aim was to compare the reproducibility of eGFR and its response to food intake with those of ⁵¹Cr-EDTA and iohexol. Finally, we exploited the availability of two gold standards to re-examine the prevailing view that eGFR is more reliable at GFR concentrations below the reference range compared with its reliability at higher concentrations of GFR.

Methods

Subjects

Subjects comprised 60 patients referred for clinical measurement of GFR (40 males, 20 females, median age 56, range 28–79 y; median body mass index [BMI] of 28, range 20–41 kg/m²) and 20 healthy volunteers (seven males, 13 females, median age 44, range 30–59 y; median BMI of 21.5, range 18–34 kg/m²). The overwhelming majority of subjects were Caucasian and none were black. Patients were asked to consider volunteering for an additional, simultaneous GFR measurement with iohexol: 36 were diabetics (21 of whom were receiving insulin), 10 had cancer, 13 had skin disease and were receiving or being considered for cyclosporine treatment and one was referred from the nephrology service. GFR was measured in the morning after the patient had consumed a routine breakfast, i.e. were non-fasting. The 20 healthy volunteers were studied twice, at the same time in the morning, once fasting and once, like the patients, after a non-standardized breakfast. Ten of the healthy volunteers had a further fasting study. Each healthy volunteer's two or three studies were completed within four weeks.

This was a prospective study completed within 18 months, during which there were no changes in laboratory creatinine practice. All subjects gave written, informed consent and the study was approved by the local research ethics committee and UK Administration of Radioactive Substances Advisory Committee.

Estimation of GFR

Plasma creatinine was measured from the baseline blood sample (see below) in all but nine subject studies (including one normal volunteer study) using the Jaffe method (Dimension^® RxL Max Clinical Chemistry analyser [Dade Behring, Milton Keynes, UK]) and eGFR calculated from a modified version of the four-variable MDRD equation.¹ The values assigned to the assay calibrators were traceable to the National Institute of Standards and Technology (NIST) Standard Reference Material (SRM) 914.

When compared with the reference method of gas chromatography/isotope dilution mass spectrometry (GC/IDMS), the bias for serum and plasma samples exceeds that stipulated by NKDEP (National Kidney Disease Education Programme).⁴⁴ In order to harmonize reported eGFR, taking into account the specificities of the assays used, it has been suggested that method-specific slope and intercept adjusters be applied to measured creatinine results.⁴⁵ As such, the calculated eGFR will be in agreement with those derived from an ‘IDMS traceable’ method using the 175 version of the equation. These adjusters have been determined from the United Kingdom National External Quality Assessment Scheme (UK NEQAS) pilot scheme for eGFR.

For dimension, intercept = 6.78 and slope = 1.030.

Hence, the calculation of eGFR using the MDRD equation appropriately modified becomes:

Correlation data, generated as part of our participation in the UKNEQAS GFR estimation scheme, between the equation above and the IDMS creatinine ‘175’ version supports this approach. Thus, ordinary least square linear regression analysis yielded the following equation (samples had creatinine concentrations in the range of 50–130 μmol/L):

The slope is not significantly different to unity (P = 0.15) and the intercept not significantly different to zero (P = 0.32).

Measurement of GFR

Two reference values of GFR were measured, one with ⁵¹Cr-EDTA and the other with iohexol (GE Healthcare, UK). The two indicators were separately injected into each antecubital fossa and assayed in contralateral samples drawn before (baseline) and 20, 40, 60, 120, 180 and 240 min after injection, ⁵¹Cr-EDTA by well-counting (Wallac 1480 Wizard 3”, Turku, Finland) and iohexol by X-ray fluorescence (Oxford Instruments, Oxford, UK), as previously described.⁴⁶ GFR was measured using a two-compartment model and scaled to a body surface area of 1.73 m².⁴⁷ In five patient studies (in one of whom a plasma creatinine measurement was also missing), the iohexol clearance curve could not be resolved into two exponentials and so they were not included in the analysis, leaving a final total of 97 subject studies for analysis.

Statistics

Taking GFR_EDTA as the primary gold standard, bias (average difference between paired measurements), imprecision (one standard deviation of bias) and disagreement (signless difference) were expressed in absolute terms and as percentages of the average. Student's unpaired t-test was used to compare biases and to measure the significance of the difference of bias from zero. Imprecisions were compared using the F-test.⁴⁸ The boundary for low and high GFR was set at 80 mL/min/1.73 m² (based on ⁵¹Cr-EDTA) as this, in European guidelines, is the minimum for live kidney donation.⁴⁹

Results

Reproducibility of eGFR

As we have previously shown by comparing the reproducibilities of GFR_EDTA and GFR_iohexol, fasting GFR changes in normal subjects between repeat studies by several percent, even under identical conditions.⁴⁶ Changes in GFR_EDTA correlated with the corresponding changes in GFR_iohexol, indicating that they were not artefactual.⁴⁶ In these same normal subjects who underwent two fasting studies, eGFR was significantly less reproducible than GFR_EDTA and GFR_iohexol (Table 1). Nevertheless, the individual changes in eGFR between the two studies correlated positively and significantly with the corresponding changes in both GFR_EDTA (r = 0.73, n = 9, P < 0.05) and GFR_iohexol (r = 0.87, n = 9, P < 0.01; Figure 1), confirming a non-random element in variation of true GFR detectable by eGFR.

Figure 1

Correlation between corresponding individual fractional simultaneous changes in estimated glomerular filtration rate (eGFR) and (a) GFR_EDTA (open circles) and (b) GFR_iohexol (closed circles) between fasting studies (left panel; n = 9) and in response to a non-standardized light meal (right panel; n = 19). With respect to the meal, the relation based on GFR_EDTA, but not on GFR_iohexol, was still statistically significant after exclusion of the individual who displayed a y-axis value of >1.8. Lines are least squares fitted regression lines

Table 1

Reproducibility, bias and imprecision (1 SD of bias) with which the first fasting study predicted the second fasting study for all three measures of glomerular filtration rate (GFR)

	Bias (%)	Imprecision (%)
GFR_EDTA	2.9	11.3
eGFR	1.5	26.9*^,†
GFR_iohexol	−3.1	15.5

*P < 0.01 vs. GFR_EDTA

^† P < 0.05 vs. GFR_iohexol

Response to food

As previously reported,⁴⁶ the non-standardized meal in the current study produces a small increase in GFR of about 5 mL/min/1.73 m². Unlike GFR_EDTA and GFR_iohexol, there was no significant change in eGFR in response to this meal in normal subjects who underwent both fasting and non-fasting studies (−3.7 SD [21.8] mL/min/1.73 m²). As in the reproducibility study, however, individual changes correlated positively with the corresponding changes in both GFR_EDTA (r = 0.52, n = 19, P < 0.05) and GFR_iohexol (r = 0.48, n = 19, P < 0.05; Figure 1).

Prediction of GFR_EDTA

Bland-Altman plots showed that the absolute differences between GFR_EDTA and eGFR and between GFR_EDTA and GFR_ophexol increased as functions of the respective averages (Figure 2), supporting the expression of the results as relative (percentage) values, as in Cirillo's study.⁸

Figure 2

Bland-Altman plots with respect to estimated glomerular filtration rate (eGFR) and GFR_EDTA (left panel) and GFR_iohexol and GFR_EDTA (right panel)

GFR_iohexol correlated closely with GFR_EDTA, closer than the correlation between eGFR and GFR_EDTA (Figure 3, Table 2). Both correlation coefficients were higher when GFR_EDTA was <80 mL/min/1.73 m².

Figure 3

Correlations between (glomerular filtration rate) GFR_EDTA and GFR_iohexol (upper panel) and between estimated GFR (eGFR) and GFR_EDTA (lower panel). Closed circles denote GFR_EDTA < 80 mL/min/1.73 m²; open circles denote GFR_EDTA > 80 mL/min/1.73 m². Lines are identity (see Table 2)

Table 2

Correlation coefficients between GFR_EDTA and estimated GFR (eGFR) and between GFR_EDTA and GFR_iohexol irrespective of renal function and in studies, respectively, with GFR_EDTA < 80 mL/min/1.73 m² (n = 37) and >80 mL/min/1.73 m² (n = 60)

	All subjects	<80	>80
eGFR	0.84	0.92	0.41
GFR_iohexol	0.97	0.98	0.86

The relative bias of GFR_iohexol, but not of eGFR, for predicting GFR_EDTA, was significantly different from zero (Table 3). eGFR predicted GFR_EDTA with significantly less precision than GFR_iohexol. The mean disagreements between GFR_EDTA and (i) eGFR and (ii) GFR_iohexol were 15.3% and 8.6%, respectively (Table 3). Imprecisions of eGFR and GFR_iohexol were inferior at GFR values <80 mL/min/1.73 m² compared with >80 mL/min/1.73 m² but not significantly (Table 4).

Table 3

Mean difference (bias), 1 SD (imprecision) and mean disagreement (signless difference) with respect to prediction of (glomerular filtration rate) GFR_EDTA by estimated GFR and by GFR_iohexol (n = 97)

	Difference, SD (%)	Disagreement, %
eGFR	−3.1 (19.9)*	15.3
GFR_iohexol	3.4^† (10.5)*	8.6

*Precisions significantly different using F-test (P < 0.01)

^†Significantly different from zero (P < 0.01)

Table 4

Dependence of estimated glomerular filtration rate (eGFR) on filtration function. Mean % difference (±1 SD: imprecision) and % disagreement (in italics) between eGFR and GFR_EDTA and between GFR_iohexol and GFR_EDTA

	eGFR	GFR_iohexol
GFR_EDTA (mL/min/1.73 m²)
<80 (n = 37)	−6.9 (22.4) 17.2 30% 19%	−4.1 (9.8)* 7.0 8% 3%
>80 (n = 60)	−0.8 (18.1) 14.1 28% 12%	8.1 (8.0)* 9.6 3% 0%

There was no significant difference between <80 and >80 mL/min/1.73 m² with respect to the percentages of paired values (eGFR vs. GFR_EDTA or GFR_iohexol vs. GFR_EDTA) disagreeing by more than 20% and 30% (denoted in bold)

*Significantly different from zero. There were no significant differences in imprecision across filtration function boundary, but bias of GFR_iohexol was significantly higher for GFR_EDTA >80 mL/min/1.73 m²

A plot of the absolute (signless) disagreement between eGFR and GFR_EDTA against the absolute (signless) disagreement between corresponding GFR_EDTA and GFR_iohexol values would consistently yield points only on one side of the line of identity if GFR_EDTA and GFR_iohexol always agreed with each other. Points were however distributed on both sides of the line of identity (Figure 4), indicating that GFR_EDTA was not always closer to GFR_iohexol but instead occasionally closer to eGFR (in 38 of 97 studies [39%]). When this plot was restricted to studies in which GFR_EDTA exceeded 80 mL/min/1.73 m², this proportion increased to 47%, but was only 24% when GFR_EDTA was <80 mL/min/1.73 m² (P < 0.05).

Figure 4

Relations between the respective corresponding absolute (signless) differences (absolute disagreement) between estimated glomerular filtration rate (eGFR) and GFR_EDTA (y-axis) and GFR_iohexol and GFR_EDTA (x-axis) shown for subjects with GFR_EDTA of <80 mL/min/1.73 m² (left panel) and >80 mL/min/1.73 m² (right panel). In the left panel, GFR_EDTA and GFR_iohexol agreed with each other less closely than GFR_EDTA with eGFR in nine of 37 subject studies (points to the right of the line of identity), whereas the corresponding proportion in the right panel was 28 of 60 (P < 0.05)

Discussion

Statistical approaches to the evaluation of eGFR have varied widely with overlapping and sometimes confusing terminology. For example, Froissart et al.²⁰ defined precision as one SD of mean difference between estimated and reference values while Cirillo et al.⁸ defined imprecision as the ‘percentage absolute’ signless difference. The terminology and statistical approach used in the current paper is therefore similar to Froissart et al.²⁰ with respect to bias and precision and similar to Cirillo et al.⁸ with respect to disagreement.

The principal findings in this study are: first, eGFR is able to detect the true, small variations in GFR that takes place between fasting measurements, which is important as it underlines its potential value in longitudinal studies in the same patient; and secondly, eGFR performs no worse, if not better, at higher levels of filtration function. As illustrated in Figure 4, this conclusion is based on absolute disagreement as well as relative disagreement.

In view of the effect of a meal on plasma creatinine concentrations, independent of changes in GFR, an inverse relation between corresponding changes in eGFR and GFR_EDTA might have been expected instead of a positive one. The meal, however, was light and, as previously shown by a positive rather than a negative relation between corresponding changes in GFR and extracellular fluid volume in response to this meal,⁵⁰ produced a primary change in GFR rather than one in creatinine.

Most previous studies concerning eGFR have focussed on absolute (i.e. mL/min/1.73 m²), rather than relative, differences. As is evident from the Bland-Altman plots (Figure 2) and from the functional categories given by Froissart et al.²⁰ this inevitably leads to increasing bias and imprecision at higher concentrations of GFR. Nevertheless, for tailoring drug therapy in cancer, for instance, relative errors in GFR measurement are more important than absolute errors, underscoring emphasis on relative differences in GFR measurements. This is because the area under the blood cancer drug concentration-time curve determines toxicity to bone marrow. This area for a drug eliminated by glomerular filtration is equal to the dose divided by renal clearance, so if GFR is measured as 25 mL/min in a patient with a true value of 5 mL/min, an absolute difference of 20 mL/min, the drug will be over-prescribed four-fold. In contrast, the same absolute difference in a patient with a true GFR of 100 mL/min, overestimated as 120 mL/min, will lead to over-prescription by a factor of only 1.2.

Focussing on absolute rather than relative difference has contributed to the view that eGFR is unreliable when GFR approaches the reference range. Percentage precision of both eGFR and GFR_iohexol at values >80 mL/min/1.73 m² was no worse than <80 mL/min/1.73 m², but biases were significantly greater for both. The underestimation of true GFR by eGFR at higher levels of function has been shown in several previous studies^{3,7,8,10,20,30,34} and the reverse at low levels of function.^7,8,12,13,20 Nevertheless, we agree with Froissart et al.²⁰ that precision is more important than bias; so, the current study suggests that eGFR performs no better at lower values of GFR. Pearson's correlation coefficient tends to obscure this, as shown in the current study, in which we found that inspite of slightly inferior precisions, eGFR and GFR_iohexol both showed better linear correlations with GFR_EDTA at low levels of GFR than at high levels. Thus, with respect to eGFR, correlation coefficients were 0.92 and 0.41 in patients with GFR_EDTA below and above 80 mL/min/1.73 m², very similar to those of Poggio et al.¹⁰ for whom they were 0.90 and 0.36 for CKD patients and renal donors, respectively, with a cut-off of 60 mL/min/1.73 m². (Correlation analysis is probably erroneous in the current context and only included in this manuscript for the purpose of comparison with previous studies.) A performance of eGFR that is not inferior at high levels of renal function is also supported by the finding that GFR_iohexol is closer to eGFR than it is to GFR_EDTA more often at higher levels of GFR than at lower levels (Figure 4).

Although iohexol is now regarded as a valid gold standard for measuring GFR, we found, as previously reported, that GFR_iohexol tended to underestimate GFR_EDTA at high values of GFR and vice versa at low values of GFR. Iohexol protein-binding may have been the cause at high GFR and extra-renal iohexol clearance the cause at low GFR, but we have no data to confirm this.

In conclusion, eGFR was significantly less accurate than GFR_iohexol for the prediction of GFR_EDTA, less reproducible than both GFR_iohexol and GFR_EDTA and failed to detect a change in GFR resulting from a small meal. Nevertheless, the use of a second gold standard exposes the contribution of errors in the primary gold standard to the perceived inaccuracy of eGFR as a predictor of true GFR. Our data also question the view that eGFR is less reliable at higher values of filtration function than at lower values.

Footnotes

Acknowledgements

This work was supported by the Kidney Research UK. We are grateful to the participating volunteers and to Dr L Tomlinson for helpful comments.

References

Levey

, Bosch

, Lewis

, Greene

, Rogers

, Roth

. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Ann Int Med 1999;130:461–70

Levey

, Coresh

, Greene

, Using standardized serum creatinine values in the Modification of Diet in Renal Disease study equation for estimating glomerular filtration rate. Ann Intern Med 2006;145:247–54

Verhave

, Fesler

, Ribstein

, du Cailar

, Mimran

. Estimation of renal function in subjects with normal serum creatinine levels: influence of age and body mass index. Am J Kidney Dis 2005;46:233–41

Burkhardt

, Hahn

, Gretz

, Gladisch

. Bedside estimation of the glomerular filtration rate in hospitalized elderly patients. Nephron Clin Pract 2005;101:c1–8

Rigalleau

, Lasseur

, Perlemoine

Estimation of glomerular filtration rate in diabetic subjects: Cockcroft formula or Modification of Diet in Renal Disease study equation?

Diabetes Care 2005;28:838–43

Macisaac

, Tsalamandris

, Thomas

, Estimating glomerular filtration rate in diabetes: a comparison of cystatin-C- and creatinine-based methods. Diabetologia 2006;49:1686–9

Rule

, Larson

, Bergstralh

, Slezak

, Jacobsen

, Cosio

. Using serum creatinine to estimate glomerular filtration rate: accuracy in good health and in chronic kidney disease. Ann Intern Med 2004;141:929–37

Cirillo

, Anastasio

, De Santo

. Relationship of gender, age, and body mass index to errors in predicted kidney function. Nephrol Dial Transplant 2005;20:1791–8

Kuan

, Hossain

, Surman

, El Nahas

, Haylor

. GFR prediction using the MDRD and Cockcroft and Gault equations in patients with end-stage renal disease. Nephrol Dial Transplant 2005;20:2394–401

10.

Poggio

, Wang

, Greene

, Van Lente

, Hall

. Performance of the Modification of Diet in Renal Disease and Cockcroft-Gault equations in the estimation of GFR in health and in chronic kidney disease. J Am Soc Nephrol 2005;16:459–66

11.

, Zuo

, Chen

, Modified glomerular filtration rate estimating equation for Chinese patients with chronic kidney disease. J Am Soc Nephrol 2006;17:2937–44

12.

Gaspari

, Ferrari

, Stucchi

, Performance of different prediction equations for estimating renal function in kidney transplantation. Am J Transplant 2004;4:1826–35

13.

Zuo

, Ma

, Zhou

, Wang

, Xu

, Wang

. Application of GFR-estimating equations in Chinese patients with chronic kidney disease. Am J Kidney Dis 2005;45:463–72

14.

Fontseré

, Bonal

, Navarro

, A comparison of prediction equations for estimating glomerular filtration rate in adult patients with chronic kidney disease stages 4–5. Effect of nutritional status and age. Nephron Clin Pract 2006;104:c160–c8

15.

Bostom

, Kronenberg

, Ritz

. Predictive performance of renal function equations for patients with chronic kidney disease and normal serum creatinine levels. J Am Soc Nephrol 2002;13:2140–4

16.

Vervoort

, Willems

, Wetzels

. Assessment of glomerular filtration rate in healthy subjects and normoalbuminuric diabetic patients: validity of a new (MDRD) prediction equation. Nephrol Dial Transplant 2002;17:1909–13

17.

Rossing

, Rossing

, Gaede

, Pedersen

, Parving

. Monitoring kidney function in type 2 diabetic patients with incipient and overt diabetic nephropathy. Diabetes Care 2006;29:1024–30

18.

Fontseré

, Salinas

, Bonal

Are prediction equations for glomerular filtration rate useful for the long-term monitoring of type 2 diabetic patients?

Nephrol Dial Transplant 2006;21:2152–8

19.

Baskar

, Venugopal

, Holland

, Singh

. Clinical utility of estimated glomerular filtration rates in predicting renal risk in a district diabetes population. Diabetes Med 2006;23:1057–60

20.

Froissart

, Rossert

, Jacquot

, Paillard

, Houillier

. Predictive performance of the modification of diet in renal disease and Cockcroft-Gault equations for estimating renal function. J Am Soc Nephrol 2005;16:763–73

21.

Lamb

, Wood

, Stowe

, O'Riordan

, Webb

, Dalton

. Susceptibility of glomerular filtration rate estimations to variations in creatinine methodology: a study in older patients. Ann Clin Biochem 2005;42:11–8

22.

Myers

, Miller

, Coresh

, Recommendations for improving serum creatinine measurement: a report from the Laboratory Working Group of the National Kidney Disease Education Program. Clin Chem 2006;52:5–18

23.

Lamb

, Webb

, Simpson

, Coakley

, Newman

, O'Riordan

. Estimation of glomerular filtration rate in older patients with chronic renal insufficiency: is the modification of diet in renal disease formula an improvement? J Am Geriatr Soc 2003;51:1012–7

24.

Lamb

, Webb

, O'Riordan

. Using the modification of diet in renal disease (MDRD) and Cockcroft and Gault equations to estimate glomerular filtration rate (GFR) in older people. Age Ageing 2007;36:689–92

25.

Laville

, Hadj-Aissa

, Pozet

, Le Bras

, Labeeuw

, Zech

. Restrictions on use of creatinine clearance for measurement of renal functional reserve. Nephron 1998;51:233–6

26.

Preiss

, Godber

, Lamb

, Dalton

, Gunn

. The influence of a cooked-meat meal on estimated glomerular filtration rate. Ann Clin Biochem 2007;44:35–42

27.

Mahajan

, Mukhiya

, Singh

Assessing suitability for renal donation: can equations predicting glomerular filtration rate substitute for a reference method in the Indian population?

Nephron Clin Pract 2005;101:c128–33

28.

Imai

, Horio

, Iseki

, Prevalence of chronic kkidney disease (CKD) in the Japanese general population predicted by the MDRD equation modified by a Japanese coefficient. Clin Exp Nephrol 2007;11:156–63

29.

Pierrat

, Gravier

, Saunders

, Predicting GFR in children and adults: a comparison of the Cockcroft-Gault, Schwartz, and modification of diet in renal disease formulas. Kidney Int 2003;64:1425–36

30.

Stoves

, Lindley

, Barnfield

, Burniston

, Newstead

. MDRD equation estimates of glomerular filtration rate in potential living kidney donors and renal transplant recipients with impaired graft function. Nephrol Dial Transplant 2002;17:2036–7

31.

Rule

, Gussak

, Pond

, Measured and estimated GFR in healthy potential kidney donors. Am J Kidney Dis 2004;43:112–9

32.

Anavekar

, Bais

, Carney

, The Australian Creatinine Consensus Working Group. Chronic kidney disease and automatic reporting of estimated glomerular filtration rate: a position statement. Clin Biochem Rev 2005;26:81–6

33.

Saleem

, Florkowski

, George

, Woltersdorf

. Comparison of two prediction equations with radionuclide glomerular filtration rate: validation in routine use. Ann Clin Biochem 2006;43:309–13

34.

Lin

, Knight

, Hogan

, Singh

. A comparison of prediction equations for estimating glomerular filtration rate in adults without kidney disease. J Am Soc Nephrol 2003;14:2573–80

35.

Rydstrom

, Tengstrom

, Cederquist

, Ahlmen

. Measurement of glomerular filtration rate by single-injection, single-sample techniques, using ⁵¹Cr-EDTA or iohexol. Scand J Urol Nephrol 1995;29:135–9

36.

Lundqvist

, Hietala

, Karp

. Experimental studies comparing iohexol and ⁵¹Cr-EDTA for glomerular filtration rate measurements. Acta Radiol 1995;36:58–63

37.

Frennby

, Sterner

, Almen

, Chai

, Jonsson

, Mansson

. Clearance of iohexol, ⁵¹Cr-EDTA and endogenous creatinine for determination of glomerular filtration rate in pigs with reduced renal function: a comparison between different clearance techniques. Scand J Clin Lab Invest 1997;57:241–52

38.

Brandstrom

, Grzegorczyk

, Jacobsson

, Friberg

, Lindahl

, Aurell

. GFR measurement with iohexol and ⁵¹Cr-EDTA. A comparison of the two favoured GFR markers in Europe. Nephrol Dial Transplant 1998;13:1176–82

39.

Pucci

, Bandinelli

, Pilo

, Nannipieri

, Navalesi

, Penno

. Iohexol as a marker of glomerular filtration rate in patients with diabetes: comparison of multiple and simplified sampling protocols. Diabet Med 2001;18:116–20

40.

Brown

, O'Reilly

. Iohexol clearance for the determination of glomerular filtration rate in clinical practice: evidence for a new gold standard. J Urol 1991;146:675–9

41.

Gaspari

, Perico

, Ruggenenti

, Plasma clearance of nonradioactive iohexol as a measure of glomerular filtration rate. J Am Soc Nephrol 1995;6:257–63

42.

Gaspari

, Perico

, Matalone

, Precision of plasma clearance of iohexol for estimation of GFR in patients with renal disease. J Am Soc Nephrol 1998;9:310–3

43.

Erley

, Bader

, Berger

, Plasma clearance of iodine contrast media as a measure of glomerular filtration rate in critically ill patients. Crit Care Med 2001;29:1544–50

44.

Myers

, Miller

, Coresh

, Recommendations for improving serum creatinine measurement: a report from the laboratory working group of the National Kidney Disease Education Program. Clin Chem 2006;52:5–18

45.

MacKenzie

. UK NEQAS for eGFR: Guidance for Creatinine Slopes and Intercepts. 2006; UK National External Quality Assessment Service. See www.ukneqas.org.uk (last accessed 18 August 2008)

46.

Bird

, Peters

, Michell

, Peters

. Reproducibilities and responses to food intake of GFR measured with chromium-51-EDTA and iohexol simultaneously and independently in normal subjects. Nephrol Dial Transplant 2008;23:1902–9

47.

Du Bois

, Du Bois

. A formula to estimate the approximate surface area if height and weight are known. Arch Intern Med 1916;17:863

48.

Altman

. Practical Statistics for Medical Research. London: Chapman and Hall, 1991

49.

A report of the Amsterdam Forum on the care of the live kidney donor: data and medical guidelines. Transplantation 2005;79:S53–66

50.

Bird

, Peters

, Michell

, Peters

. Association between glomerular filtration rate and extracellular fluid volume in normal subjects and patients with renal impairment. Scand J Clin Lab Invest 2008;68:39–49

Evaluation of the Modification of Diet in Renal Disease equation (eGFR) against simultaneous,dual-marker multi-sample measurements of glomerular filtration rate

Abstract

Background

Methods

Results

Conclusion

Introduction

Methods

Subjects

Estimation of GFR

Measurement of GFR

Statistics

Results

Reproducibility of eGFR

Response to food

Prediction of GFREDTA

Discussion

Footnotes

Acknowledgements

References

Prediction of GFR_EDTA