Determination of optimal sampling times for a two blood sample clearance method using ⁵¹Cr-EDTA in cats

Abstract

Estimation of the glomerular filtration rate (GFR) is a useful tool in the evaluation of kidney function in feline medicine. GFR can be determined by measuring the rate of tracer disappearance from the blood, and although these measurements are generally performed by multi-sampling techniques, simplified methods are more convenient in clinical practice. The optimal times for a simplified sampling strategy with two blood samples (2BS) for GFR measurement in cats using plasma ⁵¹chromium ethylene diamine tetra-acetic acid (⁵¹Cr-EDTA) clearance were investigated. After intravenous administration of ⁵¹Cr-EDTA, seven blood samples were obtained in 46 cats (19 euthyroid and 27 hyperthyroid cats, none with previously diagnosed chronic kidney disease (CKD)). The plasma clearance was then calculated from the seven point blood kinetics (7BS) and used for comparison to define the optimal sampling strategy by correlating different pairs of time points to the reference method. Mean GFR estimation for the reference method was 3.7±2.5 ml/min/kg (mean±standard deviation (SD)). Several pairs of sampling times were highly correlated with this reference method (r²≥0.980), with the best results when the first sample was taken 30 min after tracer injection and the second sample between 198 and 222 min after injection; or with the first sample at 36 min and the second at 234 or 240 min (r² for both combinations=0.984). Because of the similarity of GFR values obtained with the 2BS method in comparison to the values obtained with the 7BS reference method, the simplified method may offer an alternative for GFR estimation. Although a wide range of GFR values was found in the included group of cats, the applicability should be confirmed in cats suspected of renal disease and with confirmed CKD. Furthermore, although no indications of age-related effect were found in this study, a possible influence of age should be included in future studies.

The measurement of glomerular filtration rate (GFR) is generally considered one of the best parameters to evaluate kidney function. ^1–3 GFR determination by urinary clearance of inulin is regarded to be the gold standard, ⁴ as inulin is excreted in urine solely by glomerular filtration, does not bind to plasma proteins, is not metabolised, and is not reabsorbed through the renal tubules. This gold standard, however, requires an intravenous constant rate infusion and catheterisation of the bladder. ⁵ Although less invasive blood sampling techniques were developed, ^6–8 inulin determination remains expensive and requires a technically difficult laboratory procedure. This created the need for GFR tracers that are more readily accessible and easy to use. Many methods have been proposed for GFR measurement using both non-radioactive (eg, exogenous creatinine, ⁹ iohexol ^5,10,11 ) and radioactive (eg, ^99mTechnetium-labelled diethylene triamine penta-acetic acid (^99mTc-DTPA), ^{12 51}chromium ethylene diamine tetra-acetic acid (⁵¹Cr-EDTA) ^13–15 ) GFR markers. Single injection methods for several tracers have been proven to be reliable, and could replace the continuous infusion methods.^{16–18 Urine sampling has been replaced with blood sampling, which is more practical to apply in clinical settings.19}

Multiple sampling protocols were developed and considered very reliable, both in human ⁴ and in veterinary medicine. ¹⁶ Where multiple blood sampling is appropriate in research settings, clinical practice urges the need for a less cumbersome method of GFR measurement, as multi-sampling techniques can be stressful or even painful, especially for feline patients. ^20,21 However, fewer plasma samples do not only imply an easy-to-access and relatively low-priced screening method for kidney function, but may also lead to possible loss in accuracy of the obtained results and a greater error in the GFR estimation. ^2,22–24

In human medicine many protocols have been developed for single plasma sampling use. ^25,26 These protocols have been shown to be sufficient for a reliable estimation of GFR and turned out to be very usable in daily hospital practice, especially in pediatrics. ^{27–29 51}Cr-EDTA is hereby considered the (radioactive) tracer of choice. ^4,30 Several algorithms for GFR determination in human medicine have been proposed, and reproducibility of these methods has been investigated. ^1,31,32

In veterinary medicine single sample methods have been investigated in pigs, ^33–35 rabbits, ³⁶ dogs, ^37–39 and recently also in cats ^37,38,40 with a variety of tracers. ⁵¹Cr-EDTA was proven to be a suitable marker for GFR estimation in cats. ⁴⁰ To simplify the multi-sample technique, a single sample (SBS) protocol was established. A high correlation with the multiple sample technique, the reference method, was found. However, the calculation of the estimated GFR using a single sample method depends on several variable parameters such as the volume of distribution of the tracer, which has to be determined empirically. Moreover, this technique is inaccurate for very low clearance values. ^25,41,42

In order to reduce the error in GFR estimation from a single sample method, several studies investigated the use of a 2BS. ^43–45 These protocols combine both a protocol that is still easy to perform and does not cause too much distress to the animal with a measurement based on two samples, making it less liable to errors caused by empirically based formulas.

This study aimed to investigate the 2BS clearance method using ⁵¹Cr-EDTA in cats. A multiple plasma sample method (seven blood samples) was used as reference method in order to obtain the optimal combination of time points for blood sampling. In a previous study the applicability of a single sample method was assessed. ⁴⁰ Based on the same set of data, a 2BS method was here derived.

Materials and Methods

Animals

Forty-six cats were included in this study (30 males, 16 females), with a mean body weight of 4.1±1.1 kg (range 2.3–6.6 kg). Twenty-seven cats (median age=13 years, range 7–16 years) were presented at the Faculty of Veterinary Medicine of Ghent University for treatment of hyperthyroidism with radioactive iodine (¹³¹I). Hyperthyroidism may influence renal haemodynamic mechanisms and increased GFR can occur. ^46,47 For this reason, hyperthyroid cats were included. Nineteen cats were euthyroid (median age=8 years, range 3–10 years) and were presented at the clinic for various reasons, eg, neutering. None of the cats showed signs of illness and were considered to be healthy adult cats. The included cats underwent a general physical examination and an abdominal ultrasound was performed. A complete blood count and measurement of total thyroxine, urea and serum creatinine concentrations was performed, and a urine sample was obtained by cystocenthesis for routine urinalysis.

The GFR measurement in the hyperthyroid cats was performed prior to the administration of radioactive ¹³¹I. Administration of any antithyroidal medication was ceased at least 3 days before GFR measurement. Food was withheld from the cats 8 h before tracer injection, and also during the study, but they had free access to water at all times.

All cats were privately owned, and following faculty guidelines they were included after informing the owners and obtaining a written consent.

⁵¹Cr-EDTA clearance test

A catheter (Medex IV catheter Optiva 2, 22G) was placed in the cephalic vein for administration of the tracer. The syringe containing the tracer dose was weighed on a high precision balance (Mettler-Toledo model AB 204-S, Laboratory & Weighing Technologies, Switzerland, balance readability: 0.1 mg, SD±0.1 mg; average dose: 4.25 MBq ⁵¹Cr-EDTA – undiluted, corresponding to a volume of ±1.15 ml; ⁵¹Cr-EDTA, GE Healthcare, Diegem, Belgium). The absolute amount of ⁵¹Cr-EDTA was less than 1 mg per injection (0.64 mg/ml ⁵¹Cr-EDTA; 3.7 MBq/ml). Due to the high specific activity of the tracer, no body weight related dose has to be calculated.

Immediately before and after injection of the tracer the catheter was rinsed with 1 ml of saline solution (0.9% NaCl). For this study, the tracer was administered with a single injection.

A multiple sampling method was used as a reference to which the 2BS method could be compared. In order to obtain the reference plasma clearance values, seven blood samples were taken at a fixed time following the tracer administration. One to 2 ml of blood was drawn from the jugular vein at 5, 15, 30, 60, 120, 180 and 240 min after the injection of the bolus of ⁵¹Cr-EDTA. The precise sampling time was registered, and the blood was transferred to an EDTA-coated plasma tube. The samples were centrifuged for 5 min (1075×g), and plasma was pipetted in aliquots. The samples were then assayed the same day or the next day, and were kept refrigerated at 4°C during this period.

For each time point 400 μl of plasma was then transferred into a counting vial. The counting vials were placed in an automated gamma well-counter. The counting protocol was set at the energy level of ⁵¹Cr-EDTA, and the activity in each sample was counted during 3 min. This yields a coefficient of variation (CV) of approximately 1% for the first samples (about 10,000 counts/min per sample), and a CV of approximately 5% for the samples taken at 4 h after administration (±400 counts/min per sample).

An empty vial (blank vial) was added to this set of seven plasma samples and was used to correct the plasma samples for background activity.

In addition to the seven plasma samples and the blank vial, a vial containing 2 ml of a standard dilution of 1/1000 of the stock solution used for injection of the patient was also counted. A new standard solution was made for each new stock of ⁵¹Cr-EDTA. This standard solution was included to determine the total injected counts (TIC), in relation to the plasma counts for the specific well-counter. This allows the expression of the injected dose in counts, the same unit as used for the expression of the activity in the plasma samples.

The dose, expressed in TIC, was calculated with the following formula:

$$\text{TIC}=\left(\text{WBI}-\text{WAI}\right)\times \text{DS}\times \text{SA}$$

• WBI=weight of the syringe before injection (in g)

• WAI=weight of the syringe after injection (in g)

• DS=dilution of the standard solution

• SA=mean activity in the standard solution sample per ml (counts/ml)

Clearance calculation, reference method: single injection, multiple samples

Plasma clearance using a single injection, multiple sample protocol was considered the reference method in this study against which the simplified method of GFR estimation was compared. A bicompartmental model is considered valid to represent a single injection clearance method, ^14,15,30,48 and following this model a time–activity curve (TAC) was plotted (time in min in x-axis, plasma activity/ml in y-axis). A bi-exponential function was fitted on the plasma samples obtained at seven points in time after the bolus injection of the tracer.

The estimated GFR is then given by the dose, divided by the area under the curve (AUC):

$$\text{GFR}=\text{dose}/\left(\int \left(S_1\times e^{-T\times {\lambda }_1}+S_2\times e^{-T\times {\lambda }_2}\right)\right)=\text{dose}/\left(S_1/{\lambda }_1+S_2/{\lambda }_2\right)$$

• T=time

• S₁=intersection of the first rapid exponential with the Y-axis

• λ₁=the slope of the first rapid exponential

• S₂=intersection of the slow exponential (late exponential) with the Y-axis

• λ₂=the slope of the late exponential

GFR is expressed in ml/min, and when normalised to body weight, expressed in ml/min/kg.

Clearance calculation, single injection, two-sample method

The feasibility of estimation of ⁵¹Cr-EDTA clearance based on 2BS consisted of calculating the AUC with a one-compartment model. The plasma activity (A) was assumed to follow a mono-exponential decrease (λ) with time (T).

Then, from the injected dose and 2BS, GFR (expressed in ml/min) can be calculated as follows:

$$\text{AUC}=\underset{T=0}{\overset{\mathrm{\infty }}{\int }}{A}_0{e}^{-\lambda T}=A_0/\lambda$$

The plasma concentration A₀ at T₀ and the mono-exponential decrease λ can be calculated from plasma concentrations A₁ and A₂ (in counts/min×ml) at time T₁ and T₂:

$$A_1=A_0{e}^{-\lambda T_1}$$

$$\lambda =\ln \left(A_1/A_2\right)/\left(T_2-T_1\right)$$

AUC is then calculated from AUC=A₀/λ and GFR=dose/AUC.

In order to obtain the optimal sample time the plasma concentration (A₁) at time T₁ was calculated from the bi-exponential curve of the investigated cat, starting at 6 min after injection of the tracer, with an increment of 6 min, up to 120 min. For each T₁ several plasma concentrations (A₂) at T₂ were calculated from the bi-exponential curve of the investigated cat, starting 6 min after T₁, with an increment of 6 min, up to 204 min after T₁. For each combination of A₁ at T₁ and A₂ at T₂ GFR was calculated for the 46 cats and the correlation coefficient between the two-samples method and the reference method determined. Data calculated in this manner allowed testing for more combinations of time points, and in total over 600 combinations were tested.

The highest correlation coefficient determined the optimal time points for the two-samples method. Agreement between the results of the 7BS and 2BS methods are depicted in a Bland–Altman plot, in which the differences in results are plotted against the average of both results obtained with both techniques.

Statistical calculations were performed using Microsoft Excel.

Results

Animals

Based on the examinations before the GFR measurement, the cats included in the study were assumed to have no (overt) renal disease.

Multiple sample method

The plasma clearance of ⁵¹Cr-EDTA for the entire group of cats ranged between 0.4 and 8.7 ml/min/kg when normalised for body weight (mean±SD, 3.7±2.5 ml/min/kg).

The mean±SD for the normal cats yielded to 2.4±1.3 ml/min/kg. The mean clearance value for the hyperthyroid cats±SD was 4.3±1.9 ml/min/kg. A significant difference between the GFR of normal and hyperthyroid cats was seen (P=0.001; α=0.05). Figure 1 depicts the TAC of a hyperthyroid cat (GFR estimation 4.0 ml/min/kg); a euthyroid cat with an estimated GFR of 3.3 ml/min/kg, and a euthyroid cat with the lowest estimated GFR (0.4 ml/min/kg). The AUC extrapolated to infinity was 4.76% (SD±4.3%), indicating that sampling up to 4 h after tracer injection in this group was sufficient.

Fig 1.

TAC of a hyperthyroid cat (estimated GFR 4.0 ml/min/kg, ▪), a euthyroid cat (estimated GFR 2.3 ml/min/kg, ▴), and a euthyroid cat with low estimated GFR (estimated GFR 0.4 ml/min/kg, ♦). Time is represented in the x-axis (h), the plasma activity (in counts/min) in the y-axis.

Two-sample method

The estimated plasma clearance, calculated from several combinations for the first and second time point of blood sampling were compared to the reference method of seven blood samples.

For different combinations of time points for a two-sample protocol a high grade of correlation (r²≥0.980) with the reference method was obtained. When the first sample is taken at 24 min (0.4 h), the second sample must be taken between 126 min (2.1 h) and 204 min (3.4 h) after injection of the tracer. Very good correlation is seen also when the first sample is taken at 30 min (0.5 h) and the second samples between 132 and 234 min (2.2 and 3.9 h). The same applies to a first sample at 36 min (0.6 h) and a second sample between 168 and 240 min (2.8 and 4.0 h). For these three groups of possible combinations we obtained an r² – value of 0.980 or higher (up to 0.984).

Table 1 summarises the various possible sample combinations that yield an r² – value of 0.980 or higher.

Table 1.

The times at which the first and corresponding second blood sample can be taken, yielding an r²-value of greater than or equal to 0.980.

Time of first blood sample (min)	Time of second blood sample (min)	r2-value
24	126–138	0.980
	144–180	0.981
	186–204	0.980
30	132–138	0.980
	144–150	0.981
	156–162	0.982
	168–192	0.983
	198–222	0.984
	228–234	0.983
36	168	0.980
	174–180	0.981
	186–198	0.982
	204–228	0.983
	234–240	0.984

Figure 2 represents a Bland–Altman plot, depicting the agreement between the absolute values of the GFR estimation obtained with the 7BS and 2BS method.

Fig 2.

Bland–Altman plot of the absolute results (in ml/min/kg) of a 7BS and 2BS GFR estimation method using ⁵¹Cr-EDTA. The x-axis represents the average values of the 7BS and 2BS methods; the y-axis represents the difference between the GFR calculated with the 7BS and 2BS methods.

Discussion

The routine assessment of kidney function in veterinary medicine is currently assessed by serum analysis (urea and especially creatinine) and urine analysis (urine specific gravity). However, in experimental models using various degrees of nephrectomy, changes in serum creatinine will become obvious only when approximately 70% of the total kidney mass is removed. ⁴⁹ Both urea and serum creatinine are considered relatively insensitive parameters for the detection of early renal dysfunction, even more so as urea can be altered by non-renal influences.50

The assessment of the GFR is considered a more sensitive method for (early) detection of small changes in the kidney function. Amongst other tracers, ⁵¹Cr-EDTA, administered as a bolus injection, has proven its applicability as a GFR marker. ^13–15 Multiple sample methods are most suited in research settings, as they yield a very reliable approximation of the GFR.

In a previous study, the applicability of ⁵¹Cr-EDTA in cats has been reported based on a 7BS protocol. ⁴⁰ Estimation of the GFR was based on a bicompartmental model, as this was deemed applicable in literature, although no specific pharmacokinetic compartmental analysis was performed. ^14,15,30,48 This multiple sample method is cumbersome to perform and causes stress and discomfort to the animals. However, in clinical settings, the well-being of the patient has to be considered besides the ease of the intervention. The number of samples needed can be reduced to one or two blood samples. Although multi-sample methods give a better estimation of the actual GFR, the advantage of simplified sampling techniques lie in the gain in practicability. In order to obtain the optimal combination of time points for GFR estimation with ⁵¹Cr-EDTA using a 2BS protocol, a wide range of combinations were assessed. The time points giving an optimal r² of 0.984 are 36 min for the first sample, and 236 or 240 min for the second sample. Various combinations, however, give excellent correlations (all having an r²≥0.980), with the first sampling time between 24 and 36 min after injection of tracer, and the second sample between 120 and 240 min after injection of tracer.

Contradictory results on the number of samples in simplified methods are found in the human literature. On one hand, a 2BS method is preferred over SBS methods because of their higher correlation with multiple sample reference methods. On the other hand, single sample protocols are chosen because of the simplicity to perform and reduction in stress and discomfort to the patient – with special attention to pediatric patients. ^51,52 Despite the good correlation between the SBS method and the multiple sample method, this empirical method is not applicable in human patients in case of very low clearance values because GFR then may be overestimated.53

In veterinary medicine, comparative studies were performed between one and two-sample methods in cats and dogs using ^99mTc-DTPA and ¹³¹iodine-labelled orthoiodohippurate (¹³¹I-OIH). ^22,37 For both tracers these authors found that the results from the two-sample protocol correlated better with the 12 sample reference method (r²=0.996 for DTPA and r²=0.994 for OIH) in comparison with a single sample protocol (r²=0.951 for DTPA and r²=0.965 for OIH). Coincidentally, the best time points for a two-sample method in cats using ^99mTc-DTPA or ¹³¹I-OIH were the same, with the first sample taken at 20 min after injection and the second one at 180 min. In this study the correlation between a 2BS and the reference method (r=0.98) was also higher compared to the SBS method (r=0.94).

Concerning ⁵¹Cr-EDTA, the SBS method established in a previous study the correlation of the 2BS method with the reference multiple sample method is slightly higher when using a two-sample protocol (r²=0.984 versus 0.9414). ⁴⁰ When deciding which sampling strategy to use, the purpose of the investigation has to be taken into consideration. ^21,44,54 A single sample protocol may be sufficient for day-to-day clinical practice, where a 2BS method correlates better with the reference method, providing reliable results even in research settings, and can be used even when GFR is low. ^51,55

In feline medicine, it would be interesting to perform these studies on cats that are suspected of renal disease, in which the routinely measured blood and urine parameters are still within the normal range. A limitation of this study is the small number of cats (n=7) with low GFR values. Further validation is needed in cats with suspicion of renal disease and with clinically overt chronic kidney disease (CKD). Moreover, it would be interesting to investigate the evolution of GFR estimated with ⁵¹Cr-EDTA with progression of age, as CKD is a disease frequently seen in elderly cats.

In this study, the included cats were assumed not to have kidney disease. However, some of the GFR values were low, despite normal urea and serum creatinine levels. One of the included euthyroid cats revealed a low estimated GFR value of 0.4 ml/min/kg. Even though it is impossible to make a statement about the predictive value of this GFR estimation test due to the absence of follow-up of the renal status of the cats, this particular cat developed azotaemia and clinically overt CKD approximately 1 year after this investigation. As longitudinal follow-up on kidney function was beyond the scope of this study, it is a limitation of this study that no information was obtained for the other cats. Further research should be performed to investigate the predictive value of this test for cats with (suspicion of) CKD, and a longitudinal study should be conducted to validate the reproducibility of the proposed technique.

Conclusion

GFR estimation in the cat using ⁵¹Cr-EDTA can be performed using a 2BS method. Although further validation in cats with confirmed or suspected of CKD should be performed, GFR values obtained with a 2BS method seem to give a reliable estimation in a wide range of GFR values. Also, influence of age on plasma clearance of this marker should be investigated.