Introduction

Glomerular filtration rate (GFR) is commonly estimated with formulas based on blood creatinine concentration, such as Modification of Diet in Renal Disease (MDRD) or Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) [1, 2]. These methods are, however, too imprecise in many clinical situations which require a direct measurement of GFR with the injection of an exogenous tracer [3,4,5]. Urinary clearance of inulin is the gold standard for GFR measurement [4], but inulin was recently withdrawn from the market in France and several other countries for safety reasons [6, 7]. 51Chromium-labelled ethylenediamine tetra-acetic acid (51Cr-EDTA) has been widely used and validated against inulin as a radio-isotopic tracer for GFR measurement in Europe [8,9,10,11].

Unfortunately, the production of 51Cr-EDTA was halted for financial reasons at the end of 2018, compelling nephrologists and radio-pharmacists to urgently consider an alternative radio-isotopic method, both for initial GFR measurements and for longitudinal follow-up of patients.

Diethylenetriaminepentaacetic acid (DTPA) is very similar to EDTA and can be labelled with technetium-99 m (99mTc-DTPA) [12]. 99mTc-DTPA is mostly used for the assessment of split renal function by scintigraphy. In a French multicentre study, we recently validated the use of 99mTc-DTPA for GFR measurement after a single bolus against 51Cr-EDTA that was still available at that time [13]. Of note, one difficulty in the use of 99mTc-DTPA is the much shorter half-life of 99mTc compared to that of 51Cr (6 h vs 28 days), thus requiring the use of a correction factor of the radioactivity measurement considering the radioactive decay.

Importantly, in case of defective urine collection, which is a frequent and unpredictable issue during GFR measurement (that requires several timed urine samplings), it is not possible to use urinary clearance and we must consequently use plasma clearance. Measurement of the plasma clearance of a tracer after a single injection is based on the Bröchner-Mortensen equation, conceptualized in a bicompartmental model to predict the distribution of the tracer [14]. This model is no longer valid in case of extracellular expansion, as the distribution from plasma to extracellular fluid is distended leading to an overestimation of the GFR [5]. However, it is still possible to interpret plasma clearance using a bolus injection followed by continuous perfusion of a radiotracer in case of extracellular volume expansion, as we previously did with 51Cr-EDTA [15, 16], because it requires reaching a steady state, regardless of the volume of distribution of the tracer. The summary of the differences between these two measurement methods is presented in Table 1. Consequently, in clinical practice, to be able to give a reliable measurement of GFR for patients with extracellular volume expansion, we must therefore use a continuous infusion method. Of note, such cases are frequent because many patients requiring GFR measurement, including patients with chronic kidney disease (CKD), advanced congestive cardiac failure or ascites, have extracellular volume expansion and often encounter difficulties to completely void their bladder. Access to a GFR measurement method using continuous perfusion of a radiotracer is therefore essential in all centres previously using 51Cr-EDTA. This method has not yet been validated with 99mTc-DTPA.

Table 1 Differences between the two GFR measurement methods
Full size table

In the present work, we compared the performance of 51Cr-EDTA, just before the announced definitive withdrawal of this reference radiotracer, and 99mTc-DTPA for GFR measurement using a continuous infusion method.

Methods

Study design and participants

We conducted a prospective single centre study: all adult (age > 18 years) patients referred to our centre for GFR measurement in February and March 2019 were asked to participate, regardless of their estimated extracellular volume. Pregnancy and breastfeeding were exclusion criteria. Past medical history, treatment and anthropometric data were recorded. Extracellular compartment hyperhydration was clinically defined by the presence of oedema and/or ascites on the day of the GFR measurement. Patients with CKD were classified according to the commonly used KDIGO guidelines [17].

This study was classified as non-interventional by the DRCI (Délégation à la Recherche Clinique et à l’Innovation) of APHP (Assistance Publique–Hôpitaux de Paris). All patients received oral and written information about the study before inclusion, and signed informed consent. The study was approved by our local Ethics Committee (CERAPHP.5, IRB registration #00011928), and for scientific use, all data were anonymized. Research was conducted in accordance with good clinical practices and the Declaration of Helsinki.

Description of GFR measurements

51Cr-EDTA 3.7 MBq/mL solution for injection (GE Healthcare, France) and Technescan DTPA™ (Curium, France) radiolabelled with sodium pertechnetate (99mTc) eluate from Tekcis® generator (Curium, France) to obtain 99mTc-DTPA were the two tested radiotracers in this study. 51Cr-EDTA, which was still active and available in our unit at the time this study was conducted, was used as the reference tracer.

Injections started in the morning between 09:00 and 10:00 a.m. Fasting was not required. One bolus of each radiotracer was first administered. Then two continuous Y infusions (in order to use a single intravenous route) were concomitantly administered at a rate of 1.5 mL/min for 4 h. This simultaneous administration does not interfere with the accuracy of radiotracer determinations, due to the difference in energy of their gamma radiation. Moreover, the concomitant infusion of the two radiotracers using a Y infusion allows both tracers to be injected under the same conditions and to start clearance calculations at the same time. Administered doses were determined with calculations based on patients’ weight, body surface area and estimated GFR (using the MDRD equation [1]) (Table 2). A sample of the continuous infusion solution and a sample of the bolus syringe were taken and syringes were weighed before and after injection in order to calculate the precise injected amount of the radiotracer (necessary to calculate the infusion rate of radioactivity, using the flow rate of the infusion pump, which is fixed and known). Over four and a half hours, plasma and urine samples were collected to calculate urinary and plasma clearance as follows: after a 60 min resting period to allow the plasma concentrations of the radiotracers to reach equilibrium, urine was collected every 30 min for 8 consecutive periods and 7 blood samples were collected from the contralateral arm to the injection every 30 min, at the mid point of each urine collection period. Completion of each urine collection was checked using urine creatinine rate. One urine and plasma sample each was collected before injections in order to determine the radioactivity background in each sample matrix.

Table 2 Doses of radiotracer calculations
Full size table

Radioactivity of urinary, plasma and syringe samples were measured with the Cobra II® 5003 (Packard®) well gamma counter. Each sample was counted for 4 min in the appropriate energy window: 140–160 keV and 240–400 keV for 99mTc and 51Cr, respectively. 99mTc radioactive decay was corrected depending on when the radioactivity of each sample was counted.

Urinary clearance was calculated for each radiotracer as the average of seven clearance measurements, defined as urine activity multiplied by urine output, divided by plasma activity. Plasma clearance was calculated by dividing the rate of infused activity by the mean plasma activity at steady-state. Indeed, once the steady state is reached (i.e. stable plasma activity of the radiotracer), the infused radioactivity rate is equal to the GFR multiplied by this steady-state plasma radioactivity, according to the principle of “entries equal to outflows” [18, 19]. As intra individual measurements were compared, GFR values were not corrected for body surface area.

Statistical analysis

Measured GFR of both radiotracers were compared using paired t tests. Precision and accuracy of GFR of 99mTc-DTPA, compared with reference 51Cr-EDTA, were evaluated using bias (difference between values obtained by both radiotracers), relative bias (bias divided by value obtained using 51Cr-EDTA, expressed in percentage), intrinsic precision (absolute difference between individual bias and mean bias, divided by reference value, expressed in percentage), Pearson correlation coefficients, accuracy within 5, 10 and 30% (AW5, AW10 and AW30, percentage of 99mTc-DTPA-derived values within 5, 10 or 30% of 51Cr-EDTA-derived values, respectively), and root mean square logarithmic error (RMSLE, calculated from the difference of the logarithmic estimated and reference values). The 95% confidence intervals (CIs) for AW5, AW10, AW30 and RMSLE were calculated using 1,000 bootstrap iterations. Performance of both radiotracers was also compared graphically using linear correlation and Bland–Altman plots [20]. All tests were two-sided using a significance level of 0.05.

Results

Patients

A total of 22 patients were included in the study. Characteristics of the patients are reported in Table 3. Mean age was 63.4 ± 17.5 years, 68.2% were males and mean body mass index was 26.4 ± 4.9 kg/m2. Mean estimated GFR (using the creatinine-derived CKD-EPI equation [2]) was 55.8 ± 21.7 mL/min. Urine collection was complete (7 samples) for 17 of 22 patients, with an average of 6.5 urine collections per patient. Eleven patients had CKD: 1 (9.1%), 3 (27.3%), 1 (9.1%), 3 (27.3%) and 3 (27.3%) in stage 1, 2, 3a, 3b and 4, respectively. Patients were classified into two groups based on whether or not clinical extracellular volume expansion was present (Table 3).

Table 3 Characteristics of the study population
Full size table

GFR measurements

Linear correlations and Bland–Altman plots for urinary and plasma clearance of both radiotracers are shown in Fig. 1.

Fig. 1
figure 1

Graphical comparison of clearance values measured with 99mTc-DTPA versus those measured with 51Cr-EDTA for urinary clearance (top panels, a and b) and plasma clearance (bottom panels, c and d). Left panels (a and c) represent linear correlation between metrics obtained by 99mTc-DTPA versus those obtained with 51Cr-EDTA; the dotted lines represent identity line. Right panels (b and d) represent Bland–Altman plots (values obtained with 99mTc-DTPA minus that obtained with 51Cr-EDTA in function of average value obtained with both tracers); the full line represents mean bias while the dotted lines represent bias ± 1.96 × standard deviation. 51Cr-EDTA 51Chromium-labelled ethylene-diamine-tetra-acetic acid, 51Cr-EDTAp plasma clearance of 51Cr-EDTA, 51Cr-EDTAu urinary clearance of 51Cr-EDTA, 99mTc-DTPA Technetium-99m-diethylenetriaminepentaacetic acid, 99mTc-DTPAp plasma clearance of 99mTc-DTPA, 99mTc-DTPAu urinary clearance of 99mTc-DTPA

Full size image

Mean urinary clearance of 51Cr-EDTA and 99mTc-DTPA was 52.4 ± 22.5 and 52.8 ± 22.6 mL/min, respectively, with a mean bias of 0.39 ± 2.50 mL/min, a mean intrinsic precision of 3.56 ± 2.77%, a Pearson’s correlation coefficient of 0.994, an AW5 of 64% (95%CI 41;77) and an AW10 and AW30 of 100% (Tables 4 and 5). Sub-group analyses showed no difference for mean urinary clearance for patients with or without extracellular compartment hyperhydration (p > 0.05) (Table 4).

Table 4 Values of GFR measured using urinary and plasma clearance of 51Cr-EDTA and of 99mTc-DTPA
Full size table
Table 5 Precision and accuracy of 99mTc-DTPA urinary and plasma clearance compared to 51Cr-EDTA urinary and plasma clearance
Full size table

Mean plasma clearance of 51Cr-EDTA and 99mTc-DTPA was 54.8 ± 20.9 and 54.4 ± 20.9 mL/min, respectively, with a mean bias of − 0.43 ± 3.89 mL/min, a mean intrinsic precision of 6.24 ± 5.00%, a Pearson’s correlation coefficient of 0.983, an AW5 of 36% (95%CI 14;59), an AW10 of 77% (95%CI 59;91) and AW30 of 100% (Tables 4 and 5). Sub-group analyses showed no difference for mean plasma clearance for patients with or without extracellular compartment hyperhydration (p > 0.05) (Table 4).

Mean difference between urinary and plasma clearance was 2.38 ± 10.29 mL/min for 51Cr-EDTA and 1.56 ± 9.97 mL/min for 99mTc-DTPA (p = 0.278). For patients with extracellular compartment hyperhydration, difference was 4.05 ± 8.83 mL/min for 51Cr-EDTA and 2.99 ± 9.19 mL/min for 99mTc-DTPA (p = 0.321). For patients without extracellular compartment hyperhydration, difference was 0.39 ± 12.00 mL/min for 51Cr-EDTA and − 0.16 ± 11.07 mL/min for 99mTc-DTPA (p = 0.641).

Discussion

Our prospective comparative study showed excellent GFR measurement accuracy and precision for both urinary and plasma clearance methods using infusion protocol with 99mTc-DTPA, compared to 51Cr-EDTA.

As expected, the plasma clearance value was slightly higher than the urinary clearance value, reflecting the extra-renal clearance of the radiotracers. This difference was similar for both radiotracers (approximately 2 mL/min), confirming similar extra-renal handling of 51Cr-EDTA and 99mTc-DTPA. Note that a standard deviation of 10% of the measured clearance is considered acceptable in our centre. This margin of error of 2 mL/min can therefore be considered acceptable up to 20 mL/min of clearance. Furthermore, this same 2 mL/min error, even if proportionately larger, will generally not have clinical consequences for patients with very low GFR. It is also important to underline that, for each GFR measurement, we performed 6 to 7 urine and blood samplings in order to be able to calculate at least 6 urine clearance measurements of the reference tracer. This repetition of measures is necessary to give the result with high precision, but necessarily implies a standard deviation. Note that GFR measurements using a single point (i.e. plasma clearance of iohexol after a single injection and with a unique plasma sample) do not have standard deviation but are less precise.

We show excellent agreement between the GFR measurement based on urinary or plasma clearance obtained with 51Cr-EDTA and those obtained with 99mTc-DTPA, despite the much shorter half-life of 99mTc compared to that of 51Cr, proving that the 99mTc radioactive decay correction enables GFR measurement. Of note, this result allows comparison of 2 successive GFR measurements, even if the first was made with 51Cr-EDTA and the second with 99mTc-DTPA.

A few previous studies compared both radiotracers for GFR measurement based on a single bolus injection. In 1984, Rehling et al. showed no significant difference in plasma and urinary clearance between 99mTc-DTPA and 51Cr-EDTA in 20 patients with variable renal function [21]. Several recent studies have confirmed this result. Andersen et al. showed no clinically relevant difference in plasma clearance between 99mTc-DTPA and 51Cr-EDTA (mean bias of 1.4 mL/min) in 56 patients [22]. Simonsen et al. did not show a significant difference in mean values of GFR measured with 99mTc-DTPA or 51Cr-EDTA [23]. Moralidis et al. showed good agreement between 99mTc-DTPA and 51Cr-EDTA plasma clearance (mean bias of 0.0 mL/min, AW30 of 95%). However, urinary clearance was not measured in these 3 studies. A recent French multicentre study (including our centre) also showed excellent accuracy and precision of GFR measurement using 99mTc-DTPA for both urinary and plasma clearance methods, compared with 51Cr-EDTA, despite an approximate 2 mL/min overestimation (accuracy within 10% of 95% for the urinary clearance and 91% for the plasma clearance) [13]. However, to the best of our knowledge, the performance of 99mTc-DTPA (compared with 51Cr-EDTA) using the continuous infusion method has never been reported. Our study therefore fills this gap in the scientific literature.

Our study helps define the central role of 99mTc-DTPA, which is one of the few remaining reliable tracers for GFR measurement since 51Cr-EDTA and inulin are no longer available. Of note, iohexol is widely used with the plasma clearance method after a single injection that does not enable GFR measurement in case of extracellular volume expansion [24, 25]. Moreover, some patients may be allergic to iodinated contrast agents and iohexol cannot be used in these patients. In practice, 99mTc-DTPA (available worldwide) is the only remaining tracer allowing a valid measurement of both urinary and plasma clearance, by the continuous infusion technique, in countries where inulin has been withdrawn. Access to a continuous infusion measurement technique is mandatory, since patients with CKD often present with numerous comorbidities leading to extracellular volume expansion (making plasma clearance evaluation with a single injection unreliable) and difficulty in performing multiple urine collections (not allowing the calculation of urine clearance).

This study has many strengths. First, the prospective design of the study allows us to have no missing data, and strengthens the reliability of our results. This is also the first time that 51Cr-EDTA and 99mTc-DTPA are compared during continuous perfusion. Because all patients received both tracers, they are their own controls for the comparison of 99mTc-DTPA and 51Cr-EDTA, thus avoiding potential confounding factors. In addition, the continuous infusion technique has long been used routinely in our department, which prevents the occurrence of technical errors. Finally, as 51Cr-EDTA is no longer available, it was urgent to validate alternatives so as not to compromise the GFR measurement of patients who would benefit from it. The demonstration of very similar values obtained with 99mTc-DTPA and 51Cr-EDTA is therefore of major clinical interest.

The main limitation of our study is the limited number of included patients, but it is due to the short overlap between implementation of 99mTc-DTPA-based GFR measurement and the disruption of 51Cr-EDTA. However, the robustness of our results seems sufficient despite the small number of patients, especially because of the high precision of our measurements. The rather limited range of GFRs in the included patients can also be noted. In addition, our assessment of the extracellular compartment would have been more accurate using impedancemetry. However, the clinical assessment is consistent with our current practice, and we do not use any other method for extracellular compartment assessment prior to GFR measurement.

In conclusion, this study shows that 99mTc-DTPA is a reliable alternative to 51Cr-EDTA for GFR measurement, based on urinary or plasma clearance measurement, using a continuous infusion for patients with or without extracellular compartment hyperhydration.