**5.1 Physiological and analytical data**

24 Basic Nephrology and Acute Kidney Injury

absence of inulin excretion in two types of aglomerular fishes (goosefish, *Lophius piscatorius*  and toadfish *Osteichthyes - Lophiidae*). In the same article, Shannon measured GFR by inulin clearance in another type of fish with glomerulus, the dogfish (*Chondrichthyes* – *Squalidae*). These fishes were then treated with phlorizin which was sensed to block all tubular activity. Although the creatinine clearance in this fish was increased, the inulin clearance was not modified by this treatment (Shannon, 1934). In the same year of 1934, inulin clearance was also measured in aglomerular fish and in dogs by Richards (Richards et al., 1934). The experimentation (measuring GFR with and without phlorizin) was then repeated in man by Smith and Shannon. The results obtained in animals were confirmed in humans. Shannon was the first human who was perfused by inulin in 1935 (Shannon & Smith, 1935; Smith, 1951c). These authors had thus suggested that inulin was not secreted by renal tubules. This assertion will be thereafter confirmed by other authors with the same type of methodology (Shannon & Smith, 1935; Alving et al., 1939; Laake, 1954). Additional arguments were developed in the sixties by animal studies using micropontions in the tubules (Gutman et al., 1965). After intravenous injection, inulin is fully excreted by kidneys in urine (Shannon & Smith, 1935), even if very low concentrations of inulin are found in bile (Höber, 1930;

Inulin is doubtless the marker who has been the most investigated from a physiological point of view. In this view, it is logical that inulin is still considered as the gold standard for GFR measurement. Nevertheless, there are limitations to its use in daily practice. Because its relatively high molecular weight (5200 Da), the molecule is relatively viscous and don't quickly reach its volume of distribution. Therefore, only methods using urinary clearance with constant infusion rate seem accurate for this marker. Such methods are more cumbersome. Moreover, inulin is not easily available on the market and remains relatively costly. From our point of view, the most important limitation of inulin is the difficulty linked to its measurement in urine and plasma. Actually, several methods have been proposed and these methods are probably not interchangeable. There is no standardization in inulin measurement. We have shown that GFR results could vary from -10 to +10 mL/min in the same patient only because inulin was measured by a different method (unpublished data). Moreover, most of the methods (except the enzymatic ones) are prone to interferences with glucose measurement which is a limiting factor when measuring GFR in diabetic patients (Little, 1949). Regarding the methods for measuring inulin, we can cite the "acid" methods (Kuehnle et al., 1992; Shaffer & Somogoyi, 1933; Alving et al., 1939; Corcoran, 1952; Rolf et al., 1949; Roe, 1934; Steinitz, 1938; Hubbard & Loomis, 1942; Lentjes et al., 1994; Heyrovsky, 1956; Rolf et al., 1949), the enzymatic methods (Day & Workman, 1984; Delanghe et al., 1991; Jung et al., 1990; Summerfield et al., 1993; Dubourg et al., 2010) and the new methods by high performance liquid chromatography (HPLC) (Ruo et al., 1991; Baccard et al., 1999; Dall'Amico et al., 1995; Pastore et al., 2001). Describing these methods in detail are beyond the scope of this chapter and we propose the readers the following reference if they are interested in this

The use of inulin as GFR marker is justified by physiologic studies. The others markers that will be proposed thereafter will be compared to inulin measurements. Therefore, the use of other markers will be justified not by physiological studies (even if some

Schanker & Hogben, 1961).

topic (Delanaye et al., 2011b).

**4. Preliminary statistical considerations** 

51Cr-EDTA is an isotopic marker which has a low molecular weight (292 Da). Most of the authors consider that 51Cr-EDTA is not binding to proteins (<0,5% (Brochner-Mortensen, 1978; Bailey et al., 1970; Garnett et al., 1967; Stacy & Thorburn, 1966; Forland et al., 1966; Kempi & Persson, 1975; Forland et al., 1966)) even if Rehling described a binding to protein of 10% (Rehling et al., 1995; Rehling et al., 2001). Due to its low molecular weight, 51Cr-EDTA is freely filtrated through the glomerulus. Physiological studies about renal handling of 51Cr-EDTA are few but it seems that 51Cr-EDTA is neither secreted nor absorbed by renal tubules (Eide, 1970). This absence of secretion and absorption is also confirmed by Forland in dogs (Forland et al., 1966). Regarding the potential extra-renal excretion of 51Cr-EDTA, Garnett described a salivary and a fecal excretion under 1% in one anephric patient (Garnett et al., 1967). Brochner-Mortensen later confirmed the poor fecal excretion (less than 0.1% of the injected dose). Studying the renal excretion and the corporal global radioactivity of 8 healthy subjects after 72 hours, Brochner-Mortensen estimated that 4.5% of the 51Cr-EDTA will be retained in the body, especially in the liver and kidneys (Brochner-Mortensen et al., 1969). The difference between 51Cr-EDTA total clearance and 51Cr-EDTA urinary clearance corresponds to extra-renal clearance of the marker. With this methodology, the same authors estimated extra-renal clearance at 4 mL/min (and this extra-renal clearance remains stable for all GFR ranges)(Brochner-Mortensen & Rodbro, 1976). Jagenburg had also calculated an extra-renal clearance of 2 mL/min in two anuric dialysis patients (Jagenburg et al., 1978). Only, Rehling described a higher extra-renal clearance at 8.4% (Rehling et al., 1995).

Measurement of 51Cr-EDTA by nuclear count is very precise and easy because 51Cr-EDTA half time is long (27 days)(Chantler et al., 1969). The quantity of 51Cr-EDTA injected is

How Measuring Glomerular Filtration Rate? Comparison of Reference Methods 27

**GFR methods Statistics Results** 

Regression Correlation

Ratio Correlation BAr

Ratio Regression Correlation

Correlation Regression

Correlation Regression Ratio

Correlation Regression Ratio

Correlation Regression BAr

Correlation Regression Mean difference

Correlation 0.995

=1.075x-3.06 0.995

51Cr-EDTA underestimat es by de 14- 16%

> 1.02 0.992 1.5±8.7

0.96 ± 0.0027 =0.96x+0.26 0.994

0.974 =1.017x+1.6

0.977 =1.004x-0.032 1.004±0.013

0.91 =0.98x+6.5 0.96±0.02

0.97 =0.85x+11.42 1.5±11.7

0.984 =1.099x+4.96

6.2 mL/min

and constant infused rate

and constant infused rate

and constant infused rate

and constant infused rate

clearance and constant infused rate 51Cr-EDTA : plasmatic clearance: on 5 hours, samples every 15 min

clearance and constant infused rate

> and constant infused rate

clearance and constant infused rate 51Cr-EDTA : plasmatic clearance: samples at 5,10, 15, 20, 30, 60, 90, 120, 150, 180, 210, 240 min

clearance and constant infused rate 51Cr-EDTA : plasmatic clearance: samples timing not available

± 20 to 140 Urinary clearance

±8 to 120 Inulin: urinary

10 to 150 Urinary clearance

**(mL/min/ 1.73 m²)** 

56 NA ± 0 to 180 Urinary clearance

20 CKD 6 to 187 Urinary clearance

17 2 healthy ± 10 to 130 Inulin: urinary

21 CKD ± 10 to 160 Urinary

20 NA 6 to 166 Inulin: urinary

15 healthy 41 calcium troubles

25 Healthy and

CKD

CKD ± 0 to 150 Urinary clearance

**References Sample Population GFR range** 

39 Healthy

CKD Calcium troubles

(Garnett et al., 1967)

(Heath et al., 1968)

(Favre & Wing, 1968)

(Lavender et al., 1969)

(Brochner-Mortensen et al., 1969)

(Chantler et al., 1969)

(Stamp et al., 1970)

(Ditzel et al., 1972)

(Lingardh, 1972)

100 clearances in 28 subjects

65 clearances in 56 subjects

relatively small and therefore the irradiating dose received by the patient is very limited (absorbed dose from 0.011 to 0.0077 mSv according to the radioactive dose injected which is usually 7 MBq). This absorbed dose corresponds to the natural dose of irradiation received in one week and is much lesser than the dose received after thoracic radiography (0.02 mS). Nevertheless, we do not recommend this technique to measure GFR in pregnant women even if authors seem to use it safely (Brochner-Mortensen, 1978; Medeiros et al., 2009; Durand et al., 2006). The dose of EDTA is 1000x lesser than the dose considered as safe (Chantler et al., 1969).

#### **5.2 Clinical data**

The first studies about 51Cr-EDTA have been published in the sixties, even if studies (but with questionable methodology) had been published before with EDTA marked with 14Cr (Spencer et al., 1958; Foreman & Trujillo, 1954). In 1964, Downes was the first to give 51Cr-EDTA to cows to study the intestinal transit (Downes & Mcdonald, 1964). In 1966, Stacy and Thorburn are the first to inject 51Cr-EDTA to lambs for measuring GFR. They reported a good correlation with inulin clearance in the animal model (ratio 51Cr-EDTA/inulin was 0,95)(Stacy & Thorburn, 1966). The first scientists who will be interested in GFR measurement by 51Cr-EDTA in humans are English (Garnett et al., 1967; Favre & Wing, 1968; Garnett et al., 1967; Heath et al., 1968; Lavender et al., 1969). It must be underlined that nearly all studies published on this marker are coming from Europe because 51Cr-EDTA is not available in USA (not approved by the FDA)(Brandstrom et al., 1998). The first author who studied 51Cr-EDTA in humans is Garnett who was nuclearist in Southampton. These first data were published in The Lancet in 1967 (Garnett et al., 1967). This author injected one unique dose of 51Cr-EDTA and described a mono-exponential decrease in 51Cr-EDTA concentrations after 30 minutes. This author already evoked the plasma clearance (and the bolus injection) to measure GFR with 51Cr-EDTA. Unhopefully, Garnett did not compare his results to inulin clearance but only to creatinine clearance. However, Garnett performed and compared 56 51Cr-EDTA urinary clearances with inulin urinary clearances. He found a correlation of 0.995 and asserted that 51Cr-EDTA result were between ±5% of the inulin results which was really excellent. Thereafter, several studies were published on the same topic to compare performances of inulin clearance with urinary or plasma clearance of 51Cr-EDTA. We resumed these studies in Table 2, restricting the data to studies in adults. However, once again, the following conclusions will be drawn from studies having used the most adequate statistical methods. Globally, the performance of 51Cr-EDTA is good. Chantler, in 1969, showed that results of urinary clearance of 51Cr-EDTA was within 5% of the results of inulin (Chantler et al., 1969). This excellent concordance between urinary clearances of 51Cr-EDTA and inulin will be later confirmed by Froissart. This author showed a bias of +3 mL/min (51Cr-EDTA thus slightly overestimating inulin) and a precision of ± 4 mL/min (95% of the 51Cr-EDTA results will be + or – 8 mL/min around the bias)(Froissart et al., 2005b). The best study comparing 51Cr-EDTA plasma clearance with inulin clearance is certainly published by Medeiros in 2009 (Medeiros et al., 2009). This author showed that bias between the two GFR was 3±6 mL/min. This is one of the rare studies where accuracy 30% results are given (defined as the percentage of patients having a 51Cr-EDTA GFR within 30% of inulin GFR). Accuracy 30% for plasmatic clearance of 51Cr-EDTA is 93%. The higher performance is obtained when late blood samples (at 6 or 8 h) are considered.

relatively small and therefore the irradiating dose received by the patient is very limited (absorbed dose from 0.011 to 0.0077 mSv according to the radioactive dose injected which is usually 7 MBq). This absorbed dose corresponds to the natural dose of irradiation received in one week and is much lesser than the dose received after thoracic radiography (0.02 mS). Nevertheless, we do not recommend this technique to measure GFR in pregnant women even if authors seem to use it safely (Brochner-Mortensen, 1978; Medeiros et al., 2009; Durand et al., 2006). The dose of EDTA is 1000x lesser than the dose considered as safe

The first studies about 51Cr-EDTA have been published in the sixties, even if studies (but with questionable methodology) had been published before with EDTA marked with 14Cr (Spencer et al., 1958; Foreman & Trujillo, 1954). In 1964, Downes was the first to give 51Cr-EDTA to cows to study the intestinal transit (Downes & Mcdonald, 1964). In 1966, Stacy and Thorburn are the first to inject 51Cr-EDTA to lambs for measuring GFR. They reported a good correlation with inulin clearance in the animal model (ratio 51Cr-EDTA/inulin was 0,95)(Stacy & Thorburn, 1966). The first scientists who will be interested in GFR measurement by 51Cr-EDTA in humans are English (Garnett et al., 1967; Favre & Wing, 1968; Garnett et al., 1967; Heath et al., 1968; Lavender et al., 1969). It must be underlined that nearly all studies published on this marker are coming from Europe because 51Cr-EDTA is not available in USA (not approved by the FDA)(Brandstrom et al., 1998). The first author who studied 51Cr-EDTA in humans is Garnett who was nuclearist in Southampton. These first data were published in The Lancet in 1967 (Garnett et al., 1967). This author injected one unique dose of 51Cr-EDTA and described a mono-exponential decrease in 51Cr-EDTA concentrations after 30 minutes. This author already evoked the plasma clearance (and the bolus injection) to measure GFR with 51Cr-EDTA. Unhopefully, Garnett did not compare his results to inulin clearance but only to creatinine clearance. However, Garnett performed and compared 56 51Cr-EDTA urinary clearances with inulin urinary clearances. He found a correlation of 0.995 and asserted that 51Cr-EDTA result were between ±5% of the inulin results which was really excellent. Thereafter, several studies were published on the same topic to compare performances of inulin clearance with urinary or plasma clearance of 51Cr-EDTA. We resumed these studies in Table 2, restricting the data to studies in adults. However, once again, the following conclusions will be drawn from studies having used the most adequate statistical methods. Globally, the performance of 51Cr-EDTA is good. Chantler, in 1969, showed that results of urinary clearance of 51Cr-EDTA was within 5% of the results of inulin (Chantler et al., 1969). This excellent concordance between urinary clearances of 51Cr-EDTA and inulin will be later confirmed by Froissart. This author showed a bias of +3 mL/min (51Cr-EDTA thus slightly overestimating inulin) and a precision of ± 4 mL/min (95% of the 51Cr-EDTA results will be + or – 8 mL/min around the bias)(Froissart et al., 2005b). The best study comparing 51Cr-EDTA plasma clearance with inulin clearance is certainly published by Medeiros in 2009 (Medeiros et al., 2009). This author showed that bias between the two GFR was 3±6 mL/min. This is one of the rare studies where accuracy 30% results are given (defined as the percentage of patients having a 51Cr-EDTA GFR within 30% of inulin GFR). Accuracy 30% for plasmatic clearance of 51Cr-EDTA is 93%. The higher performance is obtained when late blood

(Chantler et al., 1969).

samples (at 6 or 8 h) are considered.

**5.2 Clinical data** 


How Measuring Glomerular Filtration Rate? Comparison of Reference Methods 29

51Cr-EDTA clearance was the first published alternative to inulin. Among the strengths of this marker, we have to underline the good performance of GFR measurement comparing to inulin (or to other markers). Physiological profile can also be considered as satisfying. This marker is yet easy to measure (especially according to its long half-life) and the precision of the measurement appears excellent. The costs, compared to other GFR markers, are acceptable. One important limitation is linked to the fact that 51Cr-EDTA GFR must be done in a Nuclear Medicine department. The most important limitation of this marker is the non-

Like 51Cr-EDTA, 99Tc-DTPA is an isotopic marker with a low molecular weight (393 Da)(Durand et al., 2006). DTPA may be labeled with another isotopic marker (113mIndium (Johansson & Falch, 1978; Reba et al., 1968; Piepsz et al., 1974), 169Ytterbium (Perrone et al., 1990; Russell et al., 1985)) but technetium 99 is the most used up to now. The 99Tc-DTPA is also used in Nuclear Imagery (isotopic nephrogram) for instance to measure separately the function or the right and left kidney (Biggi et al., 1995; Hilson et al., 1976; Kainer et al., 1979). However, we will only discuss GFR measurement based on plasma and/or urinary methods with 99Tc-DTPA. GFR can also be estimated with external counting using gamma camera (namely the "Gates" method) (Gates, 1984; Russell, 1987) but this method is not precise enough to be considered as a reference method for measuring GFR. For some authors, the GFR estimation given by the Gates method is even less performing than the creatinine clearance (Owen et al., 1982; Goates et al., 1990; van de Wiele C. et al., 1999; Ma et al., 2007; Mulligan et al., 1990; Galli et al., 1994; Ginjaume et al., 1985; Rodby et al., 1992; Tepe et al., 1987; Aydin et al., 2008; De Santo et al., 1999;

Doses of injected 99Tc-DTPA are totally safe (10 MBq)(Kempi & Persson, 1975; Durand et al., 2006). If the GFR measurement is coupled with nephrogram, the radioactive dose is however 40 to 200x higher than a simple GFR measurement with 51Cr-EDTA (Kempi & Persson, 1975; Griffiths et al., 1988). The half-life of 99Tc-DTPA is short (6.05 h) which imposes that the GFR measurement is realized quickly after the samplings, which is a practical inconvenient compared to 51Cr-EDTA (Owen et al., 1982). The 99Tc-DTPA measurement is as precise as other isotopic methods. The most relevant critic regarding 99Tc-DTPA is its potential binding to protein. This aspect has been debated in the literature. Some authors described a binding to plasma proteins from 2 to 13%, which implies an underestimation of GFR, especially when GFR is measured by plasmatic clearance (Kempi & Persson, 1975; Agha & Persson, 1977; Klopper et al., 1972; Biggi et al., 1995; Houlihan et al., 1999; Rehling et al., 2001). These high percentages could however been explained by the lack of purity of the first available preparations of 99Tc-DTPA (Rootwelt et al., 1980; Rehling et al., 2001; Fleming et al., 2004; Carlsen et al., 1980; Russell et al., 1983; Kempi & Persson, 1975). This hypothesis has been well illustrated in 1980 by Carlsen who studied and compared 51Cr-EDTA clearances with 4 different commercial preparations of 99Tc-DTPA. This author showed different results according to the preparation used (Carlsen et al., 1980). The binding to protein may also be studied by different methodologies (ultrafiltration, electrophoresis, precipitation, *in vitro* or *in vivo,* in humans or in animals etc)(Rehling et al.,

**5.3 Strengths and limitations** 

**6.1 Physiological and analytical data** 

Fawdry et al., 1985; Durand et al., 2006).

use in USA, where 51Cr-EDTA is not recognized by the FDA.

**6. 99Tc-DTPA (Diethylenetriaminepenta-acetic acid)** 


Table 2. Studies comparing 51Cr-EDTA with inulin. NA: not available, CKD: chronic kidney disease subjects, BA: Bland and Altman analysis, BAr: Bland and Altman analysis recalculated by us, BM: Brochner-Mortensen.

#### **5.3 Strengths and limitations**

28 Basic Nephrology and Acute Kidney Injury

29 CKD ± 30 to 160 Urinary clearance

31 CKD ± 30 to 160 Inulin: urinary

130 to 150 Urinary clearance

± 5 to 120 Inulin: urinary

17 Severe CKD 2.6 to 11 Urinary clearance Correlation

19 Nephrectomy 11 to 76 Inulin: urinary

111 NA NA Urinary clearance

22 NA NA Urinary clearance

44 Renal grafted ±15 to 80 Inulin: urinary

Table 2. Studies comparing 51Cr-EDTA with inulin. NA: not available, CKD: chronic kidney disease subjects, BA: Bland and Altman analysis, BAr: Bland and Altman analysis re-

and constant infused rate

and constant infused rate

clearance and constant infused rate 51Cr-EDTA : plasmatic clearance: samples at 180, 200, 220 et 240 min + BM correction

clearance and constant infused rate 51Cr-EDTA : urinary clearance:

clearance and constant infused rate 51Cr-EDTA : plasmatic clearance: 5 samples between 3 and 5 h+BM correction

> and constant infused rate

> and constant infused rate

clearance and constant infused rate 51Cr-EDTA : plasmatic clearance: samples at 2, 4, 6, 8 h + BM correction

Ratio 0.9±0.01

0.97 =0.855x+7.555 0.96±0.07

0.97 =0.961x+2.908 1±0.11

> 0.99 =0.96x+3.5

0.97 =1.05x-0.3

0.96 =0.86x+2.4 4.3 mL/min

BA 2.7±3.5

BA 4±4.9

NS 0.94 2.5±6.1 90.9%

Correlation Regression Ratio

Correlation Regression Ratio

Correlation Regression

Regression

Correlation Regression SD around the mean difference

t-test Correlation BA Exactitude 30%

(Brochner-Mortensen, 1973)

(Hagstam et al., 1974)

(Hagstam et al., 1974)

(Winterborn et al., 1977)

(Jagenburg et al., 1978)

 (Rehling et al., 1986)

(Froissart et al., 2005b)

(Froissart et al., 2005a)

(Medeiros et al., 2009)

calculated by us, BM: Brochner-Mortensen.

89 clearances in 9 subjects Healthy, before and after hyperglycemia

16 Children and 4 healthy adults

51Cr-EDTA clearance was the first published alternative to inulin. Among the strengths of this marker, we have to underline the good performance of GFR measurement comparing to inulin (or to other markers). Physiological profile can also be considered as satisfying. This marker is yet easy to measure (especially according to its long half-life) and the precision of the measurement appears excellent. The costs, compared to other GFR markers, are acceptable. One important limitation is linked to the fact that 51Cr-EDTA GFR must be done in a Nuclear Medicine department. The most important limitation of this marker is the nonuse in USA, where 51Cr-EDTA is not recognized by the FDA.
