Early detection is instrumental for slowing the progression of kidney disease, which is one of the most common causes of morbidity and death in cats.1–3 Renal function can be assessed in a variety of ways, but measuring the GFR is considered to be the best method.4–6 Measurement of the GFR provides substantial clinical advantages over measurement of serum renal parameters for early detection of kidney disease and for more accurately monitoring kidney function over time in patients.1,5,7 The GFR can be determined by a number of techniques that have advantages and disadvantages and that vary in accuracy and precision.5,7,8 Many of these techniques rely on measuring the plasma or renal clearance of a marker substance that is endogenous (eg, creatinine) or exogenous (eg, creatinine, inulin, iohexol,99mTc-DTPA [technetium Tc 99m pentetate]or EDTA labeled with chromium 51), administered IV, and only eliminated from the body by renal filtration.5,7–9 Although urinary clearance of constantly infused inulin is considered by many to be the primary reference method for evaluation of methods used to determine GFR,10 this is controversial.11,12
Clearance of an exogenous GFR marker is calculated as the GFR marker substance's mass elimination rate divided by its corresponding plasma concentration, often expressed in mL/min.7,8,13,14 Plasma clearance of a GFR marker after an IV bolus is calculated as the dose of the marker injected divided by the AUC, and various mathematical models (eg, numerical integration of plasma concentration data) have been used to calculate the AUC.7–9,15 The number of plasma samples required as well as the most appropriate timing of samples after injection of the marker substance depends on the model of analysis used.7,8 Although investigators of clinical studies16–19 have focused on protocols that limit the number of blood samples required (sometimes to only a single sample) and have compared the validity of these limited sampling methods with that of extensive sampling methods, reduced sampling techniques have been reported20 to be imprecise and to over- or underestimate more accurate GFR results.
In human medicine, GFR measurement, as determined by the Tk-GV method, has been described and validated retrospectively for 471 pediatric and adult human GFR studies in 201 patients, with and some without fluid disturbances, with 3 GFR markers including 99mTc-DTPA, diethylenetriamine pentaacetic acid labeled with ytterbium-169, and EDTA labeled with chromium-51.20–22 For radiolabeled DTPA, Tk-GV analysis was shown to be more accurate and precise than other pharmacokinetic models for plasma clearance methods used for determining GFR.21,22 The study presented here represented the first application of the Tk-GV method for determining GFR in veterinary patients.
The primary objective of the study presented here was to evaluate agreement of 4 methods (Tk-GV, plasma CrCL, GTS, and CT-Pp) for measuring GFR in healthy cats.1,4,7,23–25 Previously unexplored in cats, the more novel of these methods was Tk-GV.
Materials and Methods
Animals
Laboratory-raised, healthy adult cats were used in the study, and data were collected over a 3-month period. The study was approved by the University of Saskatchewan Animal Ethics Board and complied with the Canadian Council on Animal Care guidelines for humane animal usage.
Inclusion criteria
For inclusion, cats had to have been healthy as determined on the basis of findings from a complete physical examination, CBC, serum biochemical analyses, urinalysis, bacterial culture of urine, FeLV and FIV testing, abdominal ultrasonography, and indirect systemic blood pressure measurement (with an oscillometric device). Cats were sedated for abdominal ultrasonography, which was performed by a board-certified veterinary radiologist (JEM) and used to screen for general health, detect subclinical renal disease, and obtain individual kidney measurements.
Protocols
Housing—Cats were group housed with environmental enrichment provided during a 3-week acclimatization period immediately before the study. During the study, cats were housed individually to facilitate feeding, monitoring of appetites, and jugular catheter maintenance. The cats were also caged individually when food was withheld immediately before GFR measurements following administration of the radiopharmaceutical 99mTc-DTPA.
Diet—Dietary protein concentration and hydration status were controlled as part of the experimental design as described previously7 to mitigate variances in GFR. All GFR measurements were conducted at the same time of day (morning) to control for potential nocturnal reduction of GFR, even though cats have little diurnal variation in GFR. All cats were fed a commercial adult maintenance dry-food dieta that contained 32% dry protein on a dry-matter basis for the duration of the study. The amount fed per cat was calculated with a standard formula as follows: (BW × 30) + 70 = RER, where BW indicates the body weight in kilograms, and RER indicates the daily resting energy requirement in kilocalories. Water was available ad libitum. Body weights and body condition scores of all cats were determined and recorded before each GFR measurement session.
Multiport central venous catheter—Two days before GFR measurement and following an 8-hour period during which food was withheld, each cat underwent general anesthesia for placement of a multiport central venous catheter.b Jugular catheters were used to facilitate blood sampling and minimize stress and discomfort. Jugular catheter care was performed daily, and each port was flushed every 6 hours with 1 mL of heparinized (1,000 U/mL) sterile saline (0.9% NaCl) solution. Before each GFR measurement, food, but not water, was withheld from each cat for 12 hours, and each cat was administered a volume of lactated Ringer solution, SC, equivalent to 3% of body weight to ensure hydration and to reduce the potential influence of undetected dehydration on GFR.
Measurements of GFR—The GFR of each cat was measured with Tk-GV, plasma CrCL, GTS, and CT-Pp. To permit testing of study order interaction with the methods and to increase the number of tests performed, GFR measurements by all 4 methods were repeated in 2 sets of tests for each cat at the same time of day approximately 24 days apart, with each set of tests consisting of the Tk-GV, plasma CrCL, and GTS studies on one day and the CT-Pp studies on the next day. The order in which the GFR measurement methods were performed was not randomized because of limitations regarding handling and transporting radioactive blood samples for analysis.
Before sedation for GFR measurement, the PCV and total protein concentration were assessed for each cat, and each cat received a cephalic vein IV catheter through which exogenous creatinine and 99mTc-DTPA were administered. Cats were then sedated with ketamine hydrochloride (5 mg/kg, IV) and midazolam (0.5 mg/kg, IV) for scintigraphy. This sedation protocol was chosen because it does not affect GFR measurement in cats.26 Through a 3-way stop cock connected in line with the cephalic vein catheter, 99mTc-DTPAc (4 mCi in saline solution, IV) was administered first, then immediately followed by administration of exogenous creatinine (dry creatinine powderd [40.0 mg/kg] dissolved in 2.5 mL of sterile water) followed by a sterile saline solution bolus flush (12 mL).
The plasma CrCL was performed as previously described,27 with a slight modification of the sampling times in that 1.0-mL blood samples were collected from the jugular catheter and placed in tubes containing sodium heparin anticoagulant at 0, 5, 30, 60, 120, 240, 360, and 480 minutes after the exogenous creatinine injection. Samples were centrifuged for 10 minutes at 4°C, then the plasma was separated and stored frozen at −80°C until analyzed. The samples were shipped on ice by courier to the Michigan State University Diagnostic Center for Population and Animal Health, where creatinine concentrations were determined by use of a modified Jaffé kinetic reaction, as previously validated in cats.19 On the basis of the log-trapezoidal rule used with noncompartmental analysis method,15 the ratio of dose-to-AUC of plasma creatinine concentrations was used to calculate CrCL.
Before the 99mTc-DTPA was administered, a preinjection measurement of the radioactivity in the syringe was determined with a dose calibrator at t = 0 and recorded. The dose was then injected through the cephalic vein catheter as described.20–22 Following injection, the radioactivity in the IV catheter and syringe were measured to determine the residual dose. For Tk-GV, 1-mL blood samples were collected from the jugular catheter and placed in tubes containing sodium heparin at 5, 30, 60, 120, 240, and 360 minutes following administration of 99mTc-DTPA. The exact time of collection for each sample was recorded. Samples were centrifuged for 10 minutes at 4°C, then equal volumes of plasma from each sample were pipetted into 3 counting tubes. The plasma samples in the counting tubes were transported in a lead container to the Royal University Hospital, where a nuclear medicine technologist counted the numbers of radioactive decays of each sample for 2 minutes in a sodium iodide well counter. The decay-corrected standard prepared by the testing laboratory earlier in the day was used to cross-calibrate the cat's dose (in mCi) with the radioactivity measured in the plasma samples (in counts/mL/min). The median value resulting from the radioactive decay counting of 3 tubes for each plasma sample was then calculated so as to avoid potential pipetting volume error. This median value was converted with the cross-calibration standards into dose percentage/mL of plasma, which for the time-samples were then fit with γ variates by use of adaptive Tikhonov regularization to find the least relative error of the γ variate's rate coefficient as described elsewhere.20–22 The Tk-GV method adaptively treated AUC as an ill-posed Fredholm integral of the first kind by finding the Tikhonov smoothing factor that best reduced the error of measuring AUC, thereby mitigating underestimation of AUC in instances of fluid excess and renal failure, and optimized the robustness, precision, and accuracy of the clearance determinations.20–22
Renal scintigraphye was performed simultaneously with the initial sample collection for Tk-GV and plasma CrCL as previously described27 but with a modification in that cats were placed in dorsal recumbency as is done in human medicine. Image acquisition began at the time of IV injection of 99mTc-DTPA, and images were acquired every 6 seconds for 30 minutes. Following the procedure, cats were kept in isolation until the following morning when their GFR was measured by CT-Pp.
For CT-Pp, cats were administered ketamine (5 mg/kg, IV) and midazolam (0.5 mg/kg, IV) to effect for intubation. Once intubated, general anesthesia was maintained with isoflurane delivered in oxygen. Dynamic CTf was performed as previously described,23 but with slight modification to the dynamic acquisition to extend the duration of recording of the resulting time attenuation curve. The maximum CT scan duration and number of sequences within the dynamic acquisition were limited by the CT scanner software; thus, the CT scan protocol was tailored to maximize the time used to monitor clearance of the iohexol.g All CT scans were performed following a brief breath hold used to induce apnea and minimize motion artifact. The technical parameters of the precontrast abdominal CT images were 5-mm axial slices acquired at 120 kV, 150 mA, 0.75 seconds tube rotation time, 25 × 25-cm display field of view, 512 × 512 image matrix size, 400 HU window width, and 40 HU window level. Precontrast abdominal CT imaging was followed by dynamic CT imaging during which iohexol (0.5 mL/kg, IV) was administered by a pressure injector at a rate of 4 mL/s through the jugular catheter. For the dynamic CT images, 8-mm axial slices were acquired with all other parameters the same as were used for the CT images obtained before administration of contrast medium. Dynamic CT image acquisition occurred every second for 45 seconds, followed by a 60-second delay; then every 5 seconds for 30 seconds, followed by a 50-second delay; then every 5 seconds for 30 seconds, followed by a 50-second delay; then every 5 seconds for 30 seconds, followed by a 60-second delay; and finally, 2 last images were acquired 30 seconds apart. Image analysis and GFR calculations were made with Patlak plot analysis as described elsewhere.22 After undergoing CT, cats were returned to isolation until their gamma radiation exposure at 30 cm from the skin surface was < 5 μSv/h.
The reference value for mean GFR was set at 30.5 mL/min/kg on the basis of meta-analyses7 from which we calculated that of the methods included for measuring GFR in clinically normal cats, plasma inulin clearance (mean GFR, 30.5 mL/min/kg; 95% confidence interval, 2.67 to 3.44 mL/min/kg) had the lowest coefficient of variation.
Statistical analysis
Statistical analysis was performed with available software.h–j Data were summarized for each GFR measurement method (Tk-GV, plasma CrCL, CT-Pp, and GTS), and results were compared. Mean ± SD and median GFR measurements were reported for each method. All probabilities were expressed as 2-tailed values. For data not normally distributed, nonparametric testing of ranked gross GFR values (gross clearance of marker substance without dividing by body weight [mL/min/cat]) was used throughout, with the exception of the correlation calculations that could be inflated by ranking. Two-step cluster analysis of ranked gross GFR values to calculate the minimum Schwarz Bayesian information criterion was performed, then confirmed by general linear models repeated measures with Greenhouse-Geisser sphericity correction leading to an interaction term and study order discard, leaving only the methods themselves as contributory independent random variables. Doing so allowed for 1-way ANOVA and permitted difference of variance testing with the Levene test and testing for normality of rank-transformed gross GFR values with the Shapiro-Wilk test. Quality assurance included homogeneity of variance testing, mean ranked GFR agreement, testing for normality of clusters, and cluster mean adjusted grouped data (ie, cluster testing, including Bonferroni corrected Student t test of the selected clusters for mean differences of cluster location). Cluster analysis also allowed for ruling out the presence of GFR outliers. Measurements of GFRs from all 14 studies for each of the 4 methods were also evaluated with the paired Wilcoxon signed-rank sum test for comparing the gross GFR (mL/min/cat) results between measurement methods. Further, Pearson correlation coefficients were determined for results of GFR values obtained from the 14 studies for each of the 4 methods. For correlation analysis, GFR values were calculated as mL/min/cat, instead of mL/min/kg to avoid the introduction of weight as a common extraneous variable that could cause a mild, but spurious, increased correlation. Because the mean reference value used has been reported in units of mL/min/kg, rather than mL/min/cat, the Welch t test of departure was also used to evaluate results. Values of P < 0.05, except as otherwise listed, were considered significant.
Results
Animals
Seven healthy, sexually intact female domestic shorthair cats that had been laboratory raised were included in the study. Mean ± SD body weight of the cats was 3.65 ± 0.66 kg (range, 2.46 to 4.50 kg), and median age was 31 months (range, 17 to 47 months). All cats tested negative for FeLV and FIV, and results of hematologic assessments were unremarkable. None of the cats were azotemic. Mean ± SD concentrations for BUN and creatinine were 8.1 ± 1.7 mmol/L (reference range, 6.0 to 11.4 mmol/L) and 121.8 ± 23.2 μmol/L (reference range, 78 to 178 μmol/L), respectively. One cat had high serum concentrations of sodium (163 mmol/L; reference range, 147 to 160 mmol/L), potassium (5.8 mmol/L; reference range, 39 to 5.5 mmol/L), phosphorus (2.4 mmol/L; reference range, 1.08 to 2.21 mmol/L), and magnesium (1.47 mmol/L; reference range, 0.74 to 1.12 mmol/L), indicative of dehydration. The urine specific gravity in this cat was 1.053; however, the urine specific gravity was > 1.035 (range, 1.038 to 1.071) for all cats in the study. Further, all cats had negative results for bacterial culture of urine. Mean ± SD length of the right and left kidneys, as measured by a board-certified veterinary radiologist (JEM), were 34.9 ± 2.8 mm and 35.4 ± 3.2 mm, respectively. One cat had ultrasonographic evidence suggestive of bilateral renal infarcts but had a urine specific gravity of 1.047 as well as right and left kidney lengths of 35.1 mm and 39.1 mm, respectively.
Measurements of GFR
Each of the 4 methods used to determine GFRs in the 7 cats was performed twice, once in each of 2 sets of tests (Tk-GV, plasma CrCL, and GTS on day 1 of the test set and CT-Pp on day 2). Median time between the first and second set of tests was 18 days (range, 10 to 42 days). In total, 14 GFR measurement studies were performed for each of the 4 methods tested (Figure 1). On the initial assessment by each method, 1 cat had a median GFR (7.84 mL/min/kg) that was the highest and well above the preestablished reference mean (3.05 mL/min/kg; 95% confidence interval, 2.67 to 3.44 mL/min/kg). Although this cat had the lowest body weight (2.46 kg), it had no abnormalities detected on physical examination, no known history of medical issues, and results within reference limits for hematologic and biochemical analyses. Median GFR for this cat in the second set of tests was 4.39 mL/min/kg. Overall, median GFR measurements with Tk-GV, plasma CrCL, CT-Pp, and GTS were 2.75 mL/min/kg (range, 1.77 to 6.70 mL/min/kg), 3.14 mL/min/kg (range, 1.76 to 7.15 mL/min/kg), 2.83 mL/min/kg (range, 1.67 to 10.88 mL/min/kg), and 4.26 mL/min/kg (range, 2.87 to 8.53 mL/min/kg), respectively. Only the mean ± SD GFR measurements obtained with GTS (4.60 ± 1.63 mL/min/kg) had (P < 0.013; Bonferroni corrected Student t test) higher ranked clearance values than did measurements obtained with Tk-GV (3.11 ± 1.27 mL/min/kg), plasma CrCL (3.50 ± 1.36 mL/min/kg), and CT-Pp (3.63 ± 2.29 mL/min/kg). The least variable method was Tk-GV, with an SD of 1.27 mL/min/kg. With the Welch t test of departure, results obtained with GTS were significantly (P = 0.003) higher (51%) than the reference mean GFR (3.05 mL/min/kg); however, there were no significant departures from the reference mean GFR for results obtained with Tk-GV (2% higher, P = 0.864), plasma CrCL (15% higher, P = 0.243), or CT-Pp (19% higher, P = 0.366). Results of tests for outliers for each GFR measurement method were negative for each ranked gross GFR value (mL/min/cat).
Box-and-whisker plots of results from 14 GFR measurement studies for each of 4 GFR measurement methods (Tk-GV, plasma CrCL, CT-Pp, and GTS) used in 7 healthy adult sexually intact female domestic shorthair cats. For each plot, the lower and upper borders of the box represent the first and third quartiles, respectively; the horizontal line in the box represents the median; and the whiskers represent the range extending to the furthest observations within ± 1.5 interquartile ranges of the 1st or 3rd quartiles. The cross represents a near outlier (results between ± 1.5 to 3.0 interquartile ranges of the 1st or 3rd quartiles), and the circles represent far outliers (results outside ± 3.0 interquartile ranges of the 1st or 3rd quartiles).
Citation: American Journal of Veterinary Research 80, 4; 10.2460/ajvr.80.4.416
Two-step cluster analysis was used to assess how the ranked gross GFR values were most naturally grouped or clustered for test method and order. Four optimal clusters corresponding to the GFR measurement methods were identified, and no clustering by order of study (first vs second) was identified.
General linear model repeated measures was performed and when combined with Greenhouse-Geisser sphericity correction indicated no significant (P = 0.504 and P = 0.800, respectively) difference regarding study order or study order interaction, but did identify significant (P = 0.007) ranked GFR measurement method dependence. On the basis of this and the cluster analysis results, the noncontributing study order and interaction terms were removed, and 1-way ANOVA was performed. Results of 1-way ANOVA indicated that there were no meaningful departures from normality (of residuals) for any of the GFR measurement methods, and no meaningful inhomogeneity of variance was identified with Levene testing. For the entire group of ranked mean corrected residual GFR values, the results of Shapiro-Wilk testing for rejecting normality were not significant (P = 0.209).
When evaluated with the Wilcoxon signed rank test, the results for gross GFR measurements obtained with Tk-GV were significantly (P = 0.009 and P < 0.001, respectively) lower than those obtained with plasma CrCL and with GTS, but were not significantly (P = 0.326) different from results obtained with CT-Pp. In addition, the results obtained with plasma CrCL were significantly (P = 0.003) lower than those obtained with GTS. Further, results obtained with CT-Pp were significantly (P = 0.030) lower than those obtained with GTS, but not significantly (P = 1) different from those obtained with plasma CrCL. Of the combinations of correlations between results of methods, results for measurements obtained with Tk-GV had the highest correlations with each of the other methods and agreed most closely (r = 0.833) with results obtained with plasma CrCL, which was the next best correlation-ranked method (Table 1).
Summary results of Pearson correlation analysis of results from 14 GFR measurement studies for each of 4 GFR measurement methods (Tk-GV, plasma CrCL, CT-Pp, and GTS) used in 7 healthy adult sexually intact female domestic shorthair cats.
| Correlations | Tk-GV | Plasma CrCL | CT-Pp | GTS |
|---|---|---|---|---|
| Tk-GV | 1 | — | — | — |
| Plasma CrCL | 0.833 | 1 | — | — |
| CT-Pp | 0.493 | 0.416 | 1 | — |
| GTS | 0.741 | 0.483 | 0.317 | 1 |
Extracellular fluid volumes
Results obtained with Tk-GV also yielded data regarding extracellular fluid volumes (mean ± SD, 20.2 ± 2.9%). Unfortunately, the other methods used either did not permit extracellular fluid volume calculations to be performed at all (GTS and CT-Pp) or could have been subject to multiple reasonable objections following critical review (plasma CrCL).
Discussion
The reference range for GFR in healthy cats is poorly defined and varies depending on the measurement method and renal marker used.11 Several GFR studies performed with plasma CrCL,3,17,19,28,29 nuclear renal scintigraphy,1,25 and CT1,4,23,24,30 have been conducted in cats with varying accuracy. To our knowledge, the present study was the first application of Tk-GV for the measurement of GFR in cats, and findings indicated that measurements obtained with Tk-GV were comparable with those obtained with plasma CrCL, CT-Pp, and GTS. Further, of the 4 methods evaluated, Tk-GV yielded the tightest range of GFR measurement results (1.8 to 6.7 mL/min/kg).
Apart from renal dysfunction, other disease processes (eg, liver cirrhosis resulting in ascites, nephrotic syndrome, and cardiac disease) in animals and humans can lead to maldistribution of fluid within body compartments as well as fluid overload.21,22 These fluid disturbances can cause underestimation, but most often overestimation, of GFR by 1- or 2-compartment methods (using mono- or biexponential functions).21 In human studies20–22 of adult and pediatric patients with and without fluid disturbances (eg, pleural effusion, localized edema, or generalized fluid overload), use of Tk-GV provides GFR measurements that, unlike results with compartmental methods, are unaffected by fluid disturbance and are more accurate and precise with fewer samples and shorter required study duration. This greater accuracy has been hypothesized to be because the Tk-GV method better accounts for ongoing, prolonged redistribution of the GFR marker substance at much later times after injection than do other bolus IV methods.13 Although the present study was not designed to evaluate whether Tk-GV would also provide such advantages over other methods when used to measure GFR in cats with fluid disturbances, the results indicated that Tk-GV compared favorably with 3 other methods currently used for measuring GFR in healthy cats.
Further, not only did Tk-GV yield results comparable with those obtained with CT-Pp, plasma CrCL, and GTS in the present study, but Tk-GV also yielded GFR measurement results that had a narrower range and that were closer to the reference value than were results from the other 3 methods. In addition, results from Tk-GV were obtained from 6 samples, compared with 8 samples required for results with plasma CrCL. The GFR measurements obtained with Tk-GV and plasma CrCL were calculated from the same timed samples, and although the number of samples needed differed by only 2 samples in the present study, only 4 samples are needed with Tk-GV, thus 4 fewer samples than with plasma CrCL. This difference is clinically relevant, especially in smaller cats and cats that are less amenable to venipuncture. Protocols requiring fewer and as few as 1 plasma sample after IV bolus injection of a GFR marker have been evaluated,7,17,31 and studies that used only monoexponential models in dogs8 and humans13 (although biexponential models are recommended32 for accurate studies in humans) have used 4 or fewer samples. Nonetheless, in humans, Tk-GV provides reliable and precise results from only 4 plasma samples over 4 hours and outperforms monoexponential and biexponential models used on 7 to 16 samples.20–22 Therefore, it is likely that 4 samples can be used in animals when measuring GFR with Tk-GV. Collecting fewer blood samples from veterinary patients would be preferable because doing so is more practical in the clinical setting and can potentially reduce the overall cost associated with analysis of multiple samples for GFR measurement.11,26,33
Renal and plasma CrCL methods for measuring GFR in dogs and cats have been performed and validated, with results compared with those obtained with renal inulin clearance (the reference standard) and with plasma inulin clearance.3,5,7,28 Findings of the present study indicated that of the combinations of correlations between results of the 4 evaluated methods (Tk-GV, plasma CrCL, CT-Pp, and GTS), GFR measurements obtained with Tk-GV had the highest correlations with those of each of the other methods and best correlated with results of plasma CrCL. Although not as statistically accurate as the ranked cluster analysis, the Welch t test of departure of our results from reference value did not differ with respect to the significance of each method's mean difference from cluster analysis. The use of the Welch t test of departure was unavoidable because the available reference value used in the present study was mean GFR reported in units of mL/min/kg, whereas we measured gross GFR in units of ml/min/cat to avoid to avoid weight influencing correlation.
One cat in the present study had high measurements for GFR with all 4 methods assessed. This was similar to findings in a study3 comparing renal and plasma CrCL in healthy cats, among which 1 cat had an extremely high gross GFR (22.1 mL/min/cat). That individual cat was clinically normal with no detectable hematologic or biochemical abnormalities, similar to the single cat with high GFR measurements by all 4 methods in the present study. Reported causes for high GFRs in cats include renal hyperfunction,34 hypertension, and early thyroid disease.23 Persistent hypertension and hyperthyroidism were ruled out for the cat in the present study by results of preliminary screening; however, intermittent hypertension associated with the stress of handling could not be completely ruled out. Transient renal hyperfunction in humans has been identified as an effect of increased basal metabolic rate34 associated with protein loading from recruitment of renal reserve beyond the basal GFR state35; however, because all cats in the present study had food withheld for at least 12 hours prior to starting the GFR measurement studies, this was considered an unlikely explanation for the high GFR values in this 1 cat. In addition, when weight indexing was discarded, and gross GFR in units of mL/min/cat was used instead of GFR in units of mL/min/kg, gross GFR measurements for this single cat were the highest results for only 3 (Tk-GV, plasma CrCL, and CT-Pp) of the 4 GFR measurement methods. Moreover, GFR values in the present study tended to be of proportional error type, and, thus, the logarithms of GFR values were typically more normally distributed than the raw GFR values themselves, consistent with results reported elsewhere.33 A log-normal distribution implies occasional high GFR values that appear to be more important than they actually are. Finally, because there were no outliers after the rank transformation and data normalization for the gross GFR values in the present study, there was no compelling evidence that the single cat with high GFR values from the first GFR measurement test set represented a true outlier in the statistical sense. This suggested that in similar cases of apparent hyperfunctioning kidneys, considerable attention should be paid to avoiding confusing an occasional high GFR value from a heavy-tailed distribution of values with a meaningful departure from normality.
Two of the methods, GTS and CT-Pp, used in the present study required neither urinary nor plasma clearance for measuring GFR and have been validated in veterinary medicine.4,5,23–27,30 Renal nuclear scintigraphy with the modified GTS27 is the most commonly used method to estimate GFR in veterinary patients because of its ability to assess renal function and because of its increased sensitivity, compared with ultrasonography, for evaluation of postrenal obstruction.25 However, GTS in the present study provided the only substantially inflated results. Glomerular filtration rates > 2.5 mL/min/kg as measured with nuclear scintigraphy in healthy cats has been reported,5,25 and renal scintigraphic GFR values < 2.5 mL/min/kg have also been reported in clinically normal cats.25,26 It has also been noted that scintigraphy is less accurate than plasma clearance studies for measurement of GFR.25 There are several potential causes for this, the most important of which is that the high attenuation correction of GTS when used in humans is not applicable to smaller animals.
Gross GFR results obtained with CT-Pp had the worst agreement with the other methods in the present study. Although GFR measurements obtained with CT-Pp correlated with those obtained with Tk-GV (r = 0.493) and plasma CrCL (r = 0.416), findings through Wilcoxon signed-rank analysis indicated that only correlations between Tk-GV and plasma CrCL results, between Tk-GV and GTS results, and between plasma CrCL and GTS results for gross GFR values were significant (P = 0.009, P = 0.001, and P = 0.003, respectively). Mean ± SD GFR obtained with CT-Pp (3.63 ± 2.29 mL/min/kg) in the present study, was comparable but higher than previously reported mean GFR values (2.45 ± 0.5824 and 2.06 ± 0.62 mL/min/kg4) in healthy cats. Contrast media like iohexol can have nephrotoxic effects that can result in a transient decrease of renal blood flow and GFR.4 In the present study, Tk-GV was the only method to have yielded results with a median value lower than that obtained with CT-Pp, and the GFR marker substance used with Tk-GV was 99mTc-DTPA, which is considered an ideal GFR marker that does not adversely affect renal function.7 Further, although there was good correlation between GFR results obtained with Tk-GV, plasma CrCL, and GTS in the present study, the findings suggested that Tk-GV had the least variance.
The present study had a number of limitations. Accuracy testing of Tk-GV and the other methods assessed was not within the scope of the present study. However, GFR results obtained with Tk-GV (mean ± SD GFR, 3.11 ± 1.27 mL/min/kg) were the least variable and best agreed with the reference mean GFR (3.05 mL/min/kg7) in cats. Moreover, accuracy testing has been favorably contrasted between Tk-GV and other methods in humans.21,22 Another limitation was that the present study used only healthy female adult cats, whereas it has been suggested that variation in body size, sex, age, and breed can influence GFR,7 and there is concern that in cats, age especially may have a greater influence on the results obtained.28 Therefore, the GFR results obtained could have been specific to the present study population profile of mainly young female cats < 4 kg in body weight, compared with more diverse populations of cats with an equal number of males and females and wider distributions of ages and body weights.
The small sample size and lack of randomization were also limitations of the present study. The lack of randomization of the order of GFR measurement methods resulted from the need to limit exposure of clinicians and laboratory personnel to radioactive blood samples and limit intrahospital movements of tested cats after administration of a radioisotope. Another limitation was the use of sedation during Tk-GV, plasma CrCL, and GTS and general anesthesia for CT-Pp. However, anesthesia in veterinary medicine is essential for the evaluation of GFR by dynamic CT-Pp because the patient must be motionless at each time point with hyperventilation required.4 Therefore, the effects of general anesthesia on GFR values for CT-Pp in the present study were a consideration, and the sedation protocol used in the present study has been shown to have the least effect on GFR.26 Another limitation was that blood pressure was not measured in cats during the 15 minutes of anesthesia required for the CT procedures; therefore, changes in systemic blood pressure, as a potential cause of variation in results for the present study, could not be excluded. Further, because we had an earlier (first) and a later (second) GFR measurement study for each cat with each GFR measurement method, it was necessary to test for and exclude the influence of study order and the interaction or nesting of study order with the method type.
With Tk-GV, the method of adapting a curve fit to concentration over time to minimize the γ variate rate coefficient's relative error is a new standard for measurement of GFR in humans20–22 and appeared in the present study to have been directly applicable to measuring GFR in cats. Results of the present study indicated that GFR measurements obtained with Tk-VG were comparable with results of other more commonly used methods of GFR determination and that Tk-VG results had less variability. Further studies consisting of larger study populations of cats are needed to better evaluate the use of Tk-GV in cats and to validate the use of Tk-GV in cats and dogs.
Acknowledgments
Supported in part by the Sylvia Fedoruk Canadian Centre of Nuclear Innovation and the Western College of Veterinary Medicine's Companion Animal Health Fund.
The authors declare that there were no other conflicts of interest.
Presented in abstract form at the American College of Veterinary Internal Medicine Convention, Indianapolis, June 2015.
ABBREVIATIONS
| 99mTc-DTPA | Technetium Tc 99m diethylene triamine pentaacetic acid |
| AUC | Area under the concentration-versus-time curve |
| CrCL | Exogenous creatinine clearance |
| CT-Pp | Dynamic CT with Patlak plotting |
| GFR | Glomerular filtration rate |
| GTS | Gates gamma camera-based GFR measurement |
| Tk-GV | Tikhonov gamma variate adaptive regularization of plasma concentration-time curve fitting applied to 99mTc-DTPA plasma clearance |
Footnotes
Iams Premium Protection adult dry cat food, Iams, Dayton, Ohio.
Central venous catheter kit, MILA International Inc, Florence, Ky.
Heparin sodium injection, Hospira Healthcare Corp, Kirkland, QC, Canada.
Lactated Ringer solution, Hospira Healthcare Corp, Kirkland, QC, Canada.
GE Medical Systems, Millenium MP4, GE Healthcare, TelAviv, Israel.
Toshiba CT Aquilion 16, Toshiba Medical Systems, Markham, ON, Canada.
Omnipaque, GE Healthcare Canada Inc, Mississauga, ON, Canada.
SPSS Statistics, version 13.0, IBM Corp, Armonk, NY.
Microsoft Excel, versions 14 through 16, Microsoft Corp, Redmond, Wash.
Analyse-it for Microsoft Excel, version 2.20, Analyse-it Software Ltd, Leeds, England.
References
1. Schmidt DM, Scrivani PV, Dykes NL, et al. Comparison of glomerular filtration rate determined by use of single-slice dynamic computed tomography and scintigraphy in cats. Am J Vet Res 2012;73:463–469.
2. Lees GE. Early diagnosis of renal disease and renal failure. Vet Clin North Am Small Anim Pract 2004;34:867–885.
3. Reynolds BS, Massal MR, Nguyen P, et al. Plasma exogenous creatinine clearance in clinically healthy cats: comparison with urinary exogenous creatinine clearance, tentative reference intervals and indexation to bodyweight. Vet J 2014;202:157–165.
4. Chang J, Ahn S, Choi S, et al. Evaluation of glomerular filtration rate by use of dynamic computed tomography and Patlak analysis in clinically normal cats. Am J Vet Res 2011;72:1276–1282.
5. Kerl ME, Cook CR. Glomerular filtration rate and renal scintigraphy. Clin Tech Small Anim Pract 2005;20:31–38.
6. Heine R, Lefebvre HP. Assessment of renal function. In: Elliot J, Gauer GF, eds. BSAVA manual of canine and feline nephrology and urology. 2nd ed. Gloucester, England: British Small Animal Veterinary Association, 2007;117–125.
7. Von Hendy-Willson VE, Pressler BM. An overview of glomerular filtration rate testing in dogs and cats. Vet J 2011;188:156–165.
8. Heiene R, Moe L. Pharmacokinetic aspects of measurement of glomerular filtration rate in the dog: a review. J Vet Intern Med 1998;12:401–414.
9. Linnetz EH, Graves TK. Glomerular filtration rate in small animal practice. Compend Contin Educ Vet 2010;32:E1–E5.
10. Finco DR. Measurement of glomerular filtration rate via urinary clearance of inulin and plasma clearance of technetium Tc 99m pentetate and exogenous creatinine in dogs. Am J Vet Res 2005;66:1046–1055.
11. Finch N. Measurement of glomerular filtration rate in cats: methods and advantages over routine markers of renal function. J Feline Med Surg 2014;16:736–748.
12. Rose GA. Measurement of glomerular filtration rate by inulin clearance without urine collection. Br Med J 1969;2:91–93.
13. Wesolowski CA, Wesolowski MJ, Babyn PS, et al. Time varying apparent volume of distribution and drug half-lives following intravenous bolus injections. PLoS One 2016;11:e0158798.
14. Blaufox MD, Aurell M, Bubeck B, et al. Report of the Radionuclides in Nephrourology Committee on renal clearance. J Nucl Med 1996;37:1883–1890.
15. Purves RD. Optimum numerical integration methods for estimation of area-under-the-curve (AUC) and area-under-the-moment-curve (AUMC). J Pharmacokinet Biopharm 1992;20:211–226.
16. Miyagawa Y, Tamkemura N, Hirose H. Evaluation of single sampling method for estimation of plasma iohexol clearance in dogs and cats with various functions. J Vet Med Sci 2010;72:271–278.
17. Heiene R, Reynolds BS, Bexfield NH, et al. Estimation of glomerular filtration rate via 2- and 4-sample plasma clearance of iohexol and creatinine in clinically normal cats. Am J Vet Res 2009;70:176–185.
18. Watson ADJ, Lefebvre HP, Concordet D, et al. Plasma exogenous creatinine clearance test in dogs: comparison with other methods and proposed limited sampling strategy. J Vet Intern Med 2002;16:22–33.
19. Miyamoto K. Evaluation of single-injection method of inulin and creatinine as a renal function test in normal cats. J Vet Med Sci 1998;60:327–332.
20. Wesolowski CA, Ling L, Xirouchakis E, et al. Validation of Tikhonov adaptively regularized gamma variate fitting with 24-h plasma clearance in cirrhotic patients with ascites. Eur J Nucl Med Mol Imaging 2011;38:2247–2256.
21. Wanasundara SN, Wesolowski MJ, Barnfield MC, et al. Accurate and precise plasma clearance measurement using four 99m Tc-DTPA plasma samples over 4h. Nucl Med Commun 2016;37:79–86.
22. Wesolowski CA, Puetter RC, Ling L, et al. Tikhonov adaptively regularized gamma variate fitting to assess plasma clearance of inert renal markers. J Pharmacokinet Pharmacodyn 2010;37:435–474.
23. Granger LA, Armbrust LJ, Rankin DC, et al. Estimation of glomerular filtration rate in healthy cats using single-slice dynamic CT and Patlak plot analysis. Vet Radiol Ultrasound 2012;53:181–188.
24. Alexander K, Dunn M, Carmel EN, et al. Clinical application of Patlak plot CT-GFR in animals with upper urinary tract disease. Vet Radiol Ultrasound 2010;51:421–427.
25. Tyson R, Daniel GB. Renal scintigraphy in veterinary medicine. Semin Nucl Med 2014;44:35–46.
26. Winter MD, Miles KG, Riedesel DH. Effects of sedation protocol on glomerular filtration rate in cat as determined by use of quantitative renal scintigraphy. Am J Vet Res 2011;72:1222–1225.
27. Gates GF. Glomerular filtration rate: estimation from fractional renal accumulation of 99mTc-DTPA (stannous). AJR Am J Roentgenol 1982;138:565–570.
28. van Hoek I, Vandermeulen E, Dachateau L, et al. Comparison and reproducibility of plasma clearance of exogenous creatinine, exo-iohexol, endo-iohexol, and 51Cr-EDTA in young adult and aged healthy cats. J Vet Intern Med 2007;21:950–958.
29. van Hoek I, Lefebvre HP, Kooistra HS, et al. Plasma clearance of exogenous creatinine, exo-iohexol, and endo-iohexol in hyperthyroid cats before and after treatment with radioiodine. J Vet Intern Med 2008;22:879–885.
30. Chang J, Kin S, Jung J, et al. Assessment of glomerular filtration rate with dynamic computed tomography in normal Beagle dogs. J Vet Sci 2011;12:393–399.
31. Katayama M, Saito J, Katayama R, et al. A single-blood-sample method using inulin for estimating feline glomerular filtration rate. J Vet Intern Med 2013;27:17–21.
32. Fleming JS, Zivanovic MA, Blake GM, et al. Guidelines for the measurement of glomerular filtration rate using plasma sampling. Nucl Med Commun 2004;25:759–769.
33. Buclin T, Pechare-Bertschi A, Séchaud R, et al. Sinistrin clearance for determination of glomerular filtration rate: a reappraisal of various approaches using a new analytical method. J Clin Pharmacol 1997;37:679–692.
34. Compher C, Frankenfield D, Keim N, et al. Best practice methods to apply to measurement of resting metabolic rate in adults: a systemic review. J Am Diet Assoc 2006;106:881–903.
35. Bosch JP, Lauer A, Glabman S. Short-term protein loading in assessment of patients with renal disease. Am J Med 1984;77:873–879.
