Department of Clinical Sciences and UMR181 de Physiopathologie et Toxicologie Expérimentales INRA, ENVT, École Nationale Vétérinaire de Toulouse, 23 Chemin des Capelles, BP 87614, 31076 Toulouse Cedex 3, France, and
Results of this study were partially presented at the 25th annual ACVIM Forum in Seattle, June 6–9, 2007.
Corresponding author: Ingrid van Hoek, DVM, Department of Medicine and Clinical Biology of Small Animals, Faculty of Veterinary Medicine, University of Ghent, Salisburylaan 133, 9820 Merelbeke, Belgium; e-mail: email@example.com.
Background: Glomerular filtration rate (GFR) can be measured by clearance methods of different markers showing discrepancies and different reproducibility in healthy cats. Studies comparing different methods of GFR measurement in hyperthyroid cats have not yet been performed.
Hypothesis: Plasma clearance of exogenous creatinine (PECCT), exo-iohexol (PexICT), and endo-iohexol (PenICT) could lead to differences in GFR measurement and the need to use the same clearance method when comparing GFR before and after radioiodine treatment in hyperthyroid cats.
Animals: Fifteen client-owned hyperthyroid cats.
Methods: GFR was measured 1 day before and 1, 4, 12, and 24 weeks after treatment. Intravenous injection of iohexol was followed immediately by IV injection of creatinine. Plasma creatinine was measured by an enzymatic method. Plasma endo- and exo-iohexol were measured by high-performance liquid chromatography coupled to ultraviolet detection.
Results: Globally, the 3 GFR methods resulted in significantly different (P < .001) GFR results. GFR results among the different methods were the same (P= .999) at all time points. All 3 techniques indicated decreasing GFR after 131I treatment. For each GFR technique, a significant decrease in GFR was observed between time point 0 and all other time points. This decrease stabilized 4 weeks after treatment, with very little decline afterward.
Conclusion and Clinical Importance: It is mandatory to use the same GFR technique in follow-up studies. GFR testing at 4 weeks posttreatment could allow assessment of the final renal functional loss after treatment in hyperthyroid cats.
Glomerular function can be crudely estimated by assessing circulating blood urea nitrogen and creatinine concentrations or more precisely evaluated by estimating glomerular filtration rate (GFR). Measurement of GFR allows detection of decreased kidney function at an early stage of kidney disease (International Renal Interest Society [IRIS] stage I) before insufficiency develops (IRIS stage II or higher).1,2 The traditional gold standard for GFR measurement is urinary clearance of inulin. However, only a few studies on urinary clearance of inulin in cats have been performed.3–5 In cats, the urinary clearance of exogenous creatinine is comparable to the urinary clearance of inulin.6
Plasma clearance techniques have the advantage of being less laborious and easier to apply in a clinical setting than do urinary clearance techniques. GFR measurement by plasma clearance of iohexol (plasma iohexol clearance test, PICT) provides comparable results as does urinary clearance of exogenous creatinine in healthy cats and dogs7,8 and is useful for detection of renal dysfunction in cats.9 With high performance liquid chromatography (HPLC), both stereoisomers of iohexol (ie, exo- and endo-iohexol) can be measured. For this reason, use of HPLC allows determination of the plasma clearance of both exo- and endo-iohexol (PexICT and PenICT, respectively), thereby providing 2 measures of GFR after iohexol administration.10 However, HPLC is expensive and not readily available in veterinary practice.
The plasma clearance of exogenous creatinine (plasma exogenous creatinine clearance test, PECCT) seems to be a promising alternative for GFR measurement in cats. The PECCT is less complicated and does not require specialized equipment, use of radionuclides, or anesthesia,11 and therefore can be used in clinical practice.12 Combined use of creatinine and iohexol in a plasma exogenous creatinine–iohexol clearance test (PEC–ICT) also has been described in cats,12,13 and allows determination of different measures of GFR with minimal stress for the animals and minimal time- and space-related variation among the methods.
Cardiac output (CO) is increased in thyrotoxicosis because of positive chronotropic and inotropic effects, decreased systemic vascular resistance, and activation of the renin–angiotensin–aldosterone system.14 Autoregulatory mechanisms in the kidney of healthy animals would counteract these changes, thereby maintaining normal GFR. In thyrotoxicosis, however, intrarenal vasodilatation occurs, which, combined with increased CO, causes increased renal blood flow, glomerular hydrostatic pressure, and GFR.15 Autoregulatory mechanisms in the kidney that respond to the increased sodium and chloride reabsorption in the tubules caused by the thyrotoxicosis lead to an additional increase in GFR.16–18
GFR will decrease after restoring euthyroidism in hyperthyroid cats, regardless of the treatment chosen.18–20 Increased pretreatment GFR can mask underlying decreased kidney function that then is identified after treatment.18,21 The objective of this study was to evaluate 3 different GFR assessment methods (PECCT, PexICT, and PenICT) for follow-up of renal function in hyperthyroid cats before and after treatment with 131I.
Materials and Methods
Fifteen client-owned hyperthyroid cats were included in the study. Age at the time of inclusion in the study was 12.7 ± 2.2 years. Cats were studied when diagnosed with hyperthyroidism, presented for treatment with radioiodine at the faculty of veterinary medicine of Ghent University (Belgium), and 24 weeks after treatment (ie, decrease in serum total thyroxine [TT4] concentration and amelioration of clinical signs). Diagnosis of hyperthyroidism was based on clinical signs compatible with hyperthyroidism, increased TT4 serum concentration, and increased thyroidal uptake of . To assess the clinical condition of the cats, initial screening included physical and routine laboratory examinations (eg, CBC, biochemistry). GFR was measured 1 day before and 1, 4, 12, and 24 weeks after treatment with 131I. Biochemistry and measurement of TT4 were repeated 1, 4, 12, and 24 weeks after treatment. Cats maintained their original diet throughout the study period.
Cats were fasted for at least 10 hours before the start of the clearance test and fed immediately after the end of the sampling period. Water was offered ad libitum. GFR was measured by the combined clearance of exogenous creatinine, exo- and endo-iohexol (PEC-ICT), as described previously.12,13 Briefly, animals received 40 mg/kg creatinine and 64.7 mg/kg iohexol.a Creatinine was dissolved in a 0.9% sodium chloride solution.b First, iohexol then creatinine was administered IV via the cephalic vein. The dead space in the catheter was rinsed with 2 mL of 0.9% sodium chloride solution and the timer was started. Blood samples (2 mL) were taken by jugular venipuncture immediately before iohexol–creatinine administration, and at 5, 15, and 30 minutes and 1, 2, 3, 6, 8, and 10 hours after administration, placed in ethylene diamine tetraacetic acid tubes, and centrifuged. Aliquots of plasma were stored at −20 °C until assayed. Plasma creatinine concentration was measured by a validated enzymatic method.c The upper limit of quantification was 13.6 mg/dL, within- and between-day coefficients of variation (CV) were <3%, and there was a linear correlation between the theoretical and the measured concentration within quantification limits. Plasma endo- and exo-iohexol concentration was measured by a validated HPLC method with ultraviolet detection.13 The lower limit of quantification was 1.2 and 8.8 μg/mL for endo- and exo-iohexol, respectively. CV was <5% and calibration curves were linear at low and high concentration ranges. The ratios of exo-iohexol and endo-iohexol stereoisomers in the Omnipaque solution were 81.9 and 18.1%, respectively.
Pharmacokinetic analyses were performed by WinNonlin.d Plasma data were subjected to noncompartmental analysis with a statistical moment approach. The area under the plasma concentration versus time curve (AUC) was calculated by the trapezoidal rule with extrapolation to infinity.11 Plasma clearance of creatinine, endo-iohexol, and exo-iohexol was determined by dividing dose administered by AUC and indexed to body weight (BW) (mL/min/kg).
A general linear modele was used to test for differences between GFR techniques before and 1, 4, 12, and 24 weeks after treatment at a global significance level of .05, and the 3 techniques were compared pairwise at a Bonferroni-adjusted comparison-wise significance level of .017 (= .05/3). The same model was used to test for differences in GFR techniques among the time points (ie, before and 1, 4, 12, and 24 weeks after treatment) and for interactions between time point and method at a global significance level of .05. Moreover, for each GFR marker, time points before (0) and 1, 4, 12, and 24 weeks after treatment were compared pairwise at a significance level of .05.
The correlation between GFR values calculated by PexICT and PECCT, PenICT and PECCT, and PexICT and PenICT was expressed in scatter plots. The 95% confidence intervals of the slopes and intercepts of these scatter plots were calculated to evaluate the respective relative and absolute systematic errors in one or both of the clearance methods compared in the scatter plot. These errors lead to between-method differences, which can be evaluated with a Bland–Altman plot. A Bland–Altman plot was used to measure bias over the range of measured GFR values by comparison of PECCT and PexICT, PECCT and PenICT, and PexICT and PenICT. The difference between 2 GFR values by 2 methods in a cat at a specific time point before or 1, 4, 12, or 24 weeks after 131I treatment was plotted on the y-axis. The average of the GFR values of the 2 methods was plotted on the x-axis, which generates a scatter diagram.22
Results are expressed as mean ± standard deviation (SD).
BW before treatment was 4.1 ± 1.3 kg and increased to 5.3 ± 1.5 kg at 24 weeks after treatment. One cat was lost to follow-up 12 weeks after treatment owing to euthanasia because of malignant neoplasia of the pleura. Two cats failed to receive follow-up because of aggressive behavior, 12 and 24 weeks after treatment, respectively. The basal plasma creatinine concentration increased from 85 ± 34 μmol/L (1.0 ± 0.4 mg/dL) before treatment to 144 ± 49 μmol/L (1.6 ± 0.6 mg/dL) 24 weeks after treatment (reference values, 9–133 μmol/L [0.1–1.5 mg/dL]). Before and 1 week after treatment, 2 cats were azotemic. At 4, 12, and 24 weeks after treatment, the number of azotemic cats was 4, 8, and 9, respectively.
The serum TT4 concentration decreased from 104 ± 56 nmol/L before treatment to 20 ± 20 nmol/L 24 weeks after treatment (reference values, 14–45 nmol/L).
Comparison of GFR Methods
Seventy-two GFR assessments (each of them including the 3 markers) were performed. The mean ± SD and range for the PECCT, PexICT, and PenICT and the mean ± SD serum TT4 concentration at the different time points are presented in Table 1. The ratio between exo- and endo-iohexol concentration in the analyzed samples was 3.5 ± 0.9. The part of the AUC extrapolated to infinity expressed as percent of the total AUC was higher than 25% in 14/72 kinetics of creatinine clearance (range, 0.3–48%), but was below 25% in all kinetics of exo- and endo-iohexol clearance (range, 1–20 and 1–22%, respectively).
Table 1. Mean ± standard deviation (range) of glomerular filtration rate (GFR) measurements (mL/min/kg) with exo-iohexol (PexICT), endo-iohexol (PenICT), and exogenous creatinine (PECCT) in 15 hyperthyroid cats before (time point 0) and 1, 4, 12, and 24 weeks after treatment with radioiodine.
Serum TT4 concentration (nmol/L)
When the superscripts (a, b, c) are different between time points for a specific marker, a statistically significant difference is observed between the values. P values are provided in “Results.”
Globally, the 3 GFR methods resulted in significantly (P < .001) different GFR values. A statistically significant difference between mean values of PECCT and PexICT (−0.254 mL/min/kg, P<.05), PECCT and PenICT (−0.716 mL/min/kg, P<.001), and PexICT and PenICT (0.463 mL/min/kg, P<.001) was observed before and 1, 4, 12, and 24 weeks after treatment. These differences in GFR among different methods were the same (P= .999) at all time points. The scatter plots of GFR values calculated by either PexICT or PenICT versus PECCT and of GFR values calculated by PenICT versus PexICT are shown in Figures 1A, 2A, and 3A. A good correlation among the 3 methods is visible in Figures 1A, 2A, and 3A. The 95% confidence intervals for the slope and intercept, respectively, of these correlation plots are (0.825; 0.954) and (−0.158; 0.187) for 1A, (0.656; 0.799) and (−0.251; 0.132) for 2A, and (0.742; 0.867) and (−0.194; 0.109) for 3A. All correlation plots had no evidence of systematic errors (a value of 0 was included in all confidence intervals); however, all 3 correlation plots indicated a relative systematic error (value of 1 not included in confidence intervals). This relative systematic error is indicated by bias in the Bland–Altman plots. Bland–Altman comparisons of PECCT, PexICT, and PenICT are shown in Figures 1B, 2B, and 3B. Bias among clearance methods is clearly visible for comparison between PenICT and PECCT (2B) and PexICT (3B), with average GFR (along the x-axis) increasing difference among GFR measurements (along the y-axis). The bias is less clearly visible in comparison of PECCT and PexICT. Nonetheless, in all 3 plots, the majority of the measurements are spread in the area of the y-axis above 0, proving that clearance of creatinine structurally generates higher GFR values than do PexICT (1B) and PenICT (2B), and PexICT generates higher GFR values than does PenICT (3B). The highest difference was between PECCT and PenICT, which is visible in the highest limits of agreement (mean difference ±2SD) and mean difference (Fig 2B).
Evaluation of GFR After Treatment
All 3 techniques indicated decreased GFR after 131I treatment in all cats (Table 1). For each of the 3 techniques separately, there were significant differences (P < .001) in the GFR value for all time points.
There was a significant decrease in PexICT between time points 0 and 1 (−23%), 4 (−39%), 12 (−41%), and 24 weeks (−47%) after treatment (P < .01). There was also a significant −23% decrease in PexICT between 1 and 12 weeks (P= .041) and a significant −33% decrease between 1 and 24 weeks (P= .002). At 1, 4, 12, and 24 weeks after treatment, respectively, 12, 14, 14, and 13 cats indicated a decrease in GFR higher than the between-day variability of 8.3%, which has been described in aged healthy cats.13 Changes in PexICT are shown in Figure 4A.
Similarly, PenICT decreased significantly between time points 0 and 1 (−28%), 4 (−42%), 12 (−44%), and 24 (−50%) weeks after treatment (P < .01). It also decreased by 26% between 1 and 12 weeks after treatment (P= .020) and by 31% between 1 and 24 weeks after treatment (P= .001). At 1, 4, 12, and 24 weeks after treatment, respectively, 10, 13, 13, and 12 cats indicated a decrease in GFR higher than the between-day variability of 19.1%, which has been described in aged healthy cats.13 Changes in PenICT are shown in Figure 4B.
There was also a significant decrease in PECCT between time point 0 and 1 (−22%) (P= .002), 4 (−34%), 12 (−34%), and 24 (−40%) weeks (P < .001) after 131I treatment. No other statistically significant differences however, were observed among other time points. At 1, 4, 12, and 24 weeks after treatment, respectively, 8, 12, 9, and 11 cats indicated a decrease in GFR higher than the between-day variability of 21.6%, which has been described in aged healthy cats.13 Changes in PECCT are shown in Figure 4C.
One of the major findings of this study is that a statistically significant difference (P < .001) in the GFR value was observed among the 3 GFR assessment techniques, whatever the time of GFR testing (ie, before and after treatment). The difference in GFR assessment among the techniques is noted in the deviation of the slope from the value 1, but the correlation among the 3 different clearance techniques, however, seems to be acceptable (Figs 1A, 2A, and 3A). A more important factor when comparing different techniques is between-method differences as a measurement of agreement with each other. This can be shown graphically in Bland–Altman plots (Figs 1B, 2B, and 3B). The difference between PenICT and PECCT and PexICT, respectively, increased with increasing mean GFR, thereby producing a bias. This can be caused by an underestimation of PenICT, an overestimation of PECCT and PexICT respectively, or both. PECCT generated systematically higher values of GFR compared with PexICT in hyperthyroid cats before and after treatment, which is comparable to findings in healthy cats.13 However, GFR values measured with PenICT are higher compared with PECCT and PexICT in healthy cats, in contrast to the lower PenICT compared with PECCT and PexICT in the hyperthyroid cats described in this study. There is no major bias visible in the Bland–Altman graph showing the comparison between the PECCT and PexICT (Fig 1B). The mean difference between these techniques is low. In combination with the good reproducibility that is described for PECCT,13 the results of the present study suggest that the PECCT is a reasonable alternative for PexICT in hyperthyroid cats before and after treatment.
The differences in GFR values according to the technique used can first be explained by laboratory variations. Creatinine and iohexol were measured in different laboratories by different assays. Nevertheless, creatinine as well as iohexol assays used in the present study had been validated previously.13 The gold standard assay method for creatinine in plasma is HPLC. However, the enzymatic method used here is most frequently used in routine clinical laboratories and is proven to be a reasonable alternative to the HPLC method.12 Other conditions (eg, storage time and temperature of plasma samples) were similar for both iohexol and creatinine. Interestingly, a significant difference between GFR values obtained by PexICT and PenICT was observed in this study, whereas both markers were assayed by the same HPLC method in the same laboratory. Colorimetric methods,8,9 atomic emission spectroscopy,20 or X-ray fluorescence methods,23–25 measuring the amount of iodine and consequently indirectly iohexol, have also been proposed for plasma iohexol assay. Iohexol assay by HPLC has good specificity, sensitivity, accuracy, and reproducibility and, moreover, allows separate measurement of the 2 stereoisomers: endo- and exo-iohexol.26 In dogs, this latter compound, which represents the major stereoisomer, is frequently selected as a GFR marker.10,26 Discrepancies have been described between GFR calculation with exo- and endo-iohexol in dogs10 and cats,12 but not in humans.27 The reason for this difference remains unclear, but these results emphasize the fact that the dispositions of exo- and endo-iohexol are not the same and consequently assay of total iohexol could lead to misinterpretation because it is a hybrid concentration reflecting the sum of 2 stereoisomers with different clearances.
The combined use here of creatinine, exo- and endo-iohexol, as performed previously in dogs and cats,7,8,12,13 also raises the issue of a potential interference among the analytes. Such an interference, however, is unlikely. Creatinine is an endogenous compound and the peak concentration observed here (up to 1795 μmol/L [20.3 mg/dL]) could be observed in severely azotemic patients. Disposition of exo- and endo-iohexol, moreover, does not seem to be affected by mild to moderate azotemia in cats.12
In the present study, pharmacokinetic analysis also cannot explain the differences observed because a noncompartmental approach was used similarly for PECCT, PexICT, and PenICT, as described in dogs.10,11 The noncompartmental and compartmental approaches moreover are comparable for clearance of iohexol28,29 and creatinine11 in dogs. Despite the wide ranges of the AUC parts extrapolated to infinity, 80% of the pharmacokinetic analyses indicated an AUC part extrapolated to infinity lower than 25% of the complete AUC. Consequently, the sampling strategy (ie, number of blood samples and time of last sampling) can be considered appropriate in healthy, hyperthyroid, and moderately azotemic cats. In more severely azotemic cats, the last blood sample should be delayed, especially for PECCT.
Discrepancies in GFR results can also be caused by physiologic differences in renal handling of the substances.23 Because a combined PEC–ICT was used, factors related to the cats themselves cannot explain the difference in PECCT, PexICT, and PenICT. Moreover, creatinine6 and iohexol7 appear to be reliable GFR markers for cats. Use of urinary clearance of inulin, considered the gold standard method, would have been helpful to compare the accuracy of each GFR marker in the follow-up of the animals. Urinary clearance testing, however, is difficult to propose to owners, in our experience, because it is tedious and time consuming for the staff, it is stressful (eg, anesthesia may be required) and potentially harmful (eg, urinary tract infection) for the animal, and an accurate measurement of urine volume is often difficult. Clearances of both creatinine and iohexol have already been proposed as alternatives for GFR measurements in cats.7–9,12
Whatever the cause of the differences among PexICT, PenICT, and PECCT, all 3 techniques nevertheless indicated the same trend with decreasing GFR after 131I treatment. This finding indicates that if the same marker is used for repeated GFR testing, it will provide clinically relevant information similar to what would have been provided by the use of the other 2 GFR markers. For each of the 3 techniques, a significant decrease in the GFR value was observed between time point 0 and all other time points. This decrease stabilized 4 weeks after treatment, with very little decline afterward, although GFR values determined by PECCT at time point 1 week were not statistically significant from those observed at time points 4 and 24 weeks. These results indicate that PECCT is a promising alternative for GFR measurement, although it might be less sensitive to small changes in GFR compared with PexICT and PenICT. The decrease in GFR from 0 to 24 weeks after treatment was relatively consistent, irrespective of the marker used (47, 50, and 40% for PexICT, PenICT, and PECCT, respectively). This is much higher than the between-day CV of each method (8.3, 19.1, and 21.6%, respectively) in aged healthy cats13 and proves that the decrease in GFR is caused by a change in glomerular function and not by intrinsic between-day variability. Several studies in the literature have investigated GFR in hyperthyroid cats before and after treatment and describe a decrease in GFR after treatment as shown in this study. However, the follow-up in these studies was measured only before and at 1 time point (6 days, 30 days, or 6 weeks) after treatment. Hence, these follow-up studies are shorter and less extensive compared with the study described here.18–20
In conclusion, PexICT, PenICT, and PECCT, although providing different GFR values, can be used for follow-up of the decrease in GFR observed in hyperthyroid cats after treatment. Nevertheless, the same GFR marker should be used throughout the follow-up period. GFR testing at 4 weeks posttreatment could also be reasonably recommended to estimate the final loss in renal function in cats after 131I treatment. Nevertheless, further investigations in a larger population are needed.