Single-slice dynamic computed tomographic determination of glomerular filtration rate by use of Patlak plot analysis in anesthetized pigs

Kate Alexander Department of Clinical Sciences, Faculty of Veterinary Medicine, Université de Montréal, Saint-Hyacinthe, QC J2S 7C6, Canada

Search for other papers by Kate Alexander in
Current site
Google Scholar
PubMed
Close
 DMV, MSc
,
Jérôme R. E. del Castillo Department of Veterinary Biomedicine, Faculty of Veterinary Medicine, Université de Montréal, Saint-Hyacinthe, QC J2S 7C6, Canada

Search for other papers by Jérôme R. E. del Castillo in
Current site
Google Scholar
PubMed
Close
 DMV, PhD
,
Norma Ybarra Department of Veterinary Biomedicine, Faculty of Veterinary Medicine, Université de Montréal, Saint-Hyacinthe, QC J2S 7C6, Canada

Search for other papers by Norma Ybarra in
Current site
Google Scholar
PubMed
Close
 MVZ
,
Valérie Morin Department of Veterinary Biomedicine, Faculty of Veterinary Medicine, Université de Montréal, Saint-Hyacinthe, QC J2S 7C6, Canada

Search for other papers by Valérie Morin in
Current site
Google Scholar
PubMed
Close
 BSc
,
Domianique Gauvin Department of Veterinary Biomedicine, Faculty of Veterinary Medicine, Université de Montréal, Saint-Hyacinthe, QC J2S 7C6, Canada

Search for other papers by Domianique Gauvin in
Current site
Google Scholar
PubMed
Close
 BSc
,
Simon Authier Department of Veterinary Biomedicine, Faculty of Veterinary Medicine, Université de Montréal, Saint-Hyacinthe, QC J2S 7C6, Canada

Search for other papers by Simon Authier in
Current site
Google Scholar
PubMed
Close
 DMV, MSc
,
Patrick Vinay Department of Medicine, Centre Hospitalier de l'Université de Montréal–Notre-Dame Hospital, 1560 Sherbrooke St E, Montreal, QC H2L 4M1, Canada

Search for other papers by Patrick Vinay in
Current site
Google Scholar
PubMed
Close
 MD, PhD
, and
Éric Troncy Department of Veterinary Biomedicine, Faculty of Veterinary Medicine, Université de Montréal, Saint-Hyacinthe, QC J2S 7C6, Canada

Search for other papers by Éric Troncy in
Current site
Google Scholar
PubMed
Close
 DV, PhD, DUn

Abstract

Objective—To compare glomerular filtration rate (GFR) as estimated from Patlak plot analysis by use of single-slice computed tomography (CT) with that obtained from clearance of plasma inulin in pigs.

Animals—8 healthy anesthetized juvenile pigs.

Procedures—All pigs underwent precontrast, whole-kidney, helical CT; postcontrast single-slice dynamic CT; and postcontrast, whole-kidney CT for volume determination. On dynamic images, corrected Hounsfield unit values were determined for each kidney and the aorta. A Patlak plot for each kidney was generated, and plasma clearance per unit volume was multiplied by renal volume to obtain whole-animal contrast clearance. Mean GFR determined via inulin clearance (Inu-GFR) was measured from each kidney and correlated to mean GFR determined via CT (CT-GFR) for the left kidney, right kidney, and both kidneys by use of linear regression and Bland-Altman analyses.

Results—CT-GFR results from 7 pigs were valid. Total and right kidney Inu-GFR were correlated with total and right kidney CT-GFR (total, R 2 = 0.85; right kidney, R 2 = 0.86). However, left kidney CT-GFR was poorly correlated with left kidney Inu-GFR (R 2 = 0.47). Bland-Altman analysis revealed no significant bias between Inu-GFR and CT-GFR for the left kidney, right kidney, or both kidneys.

Conclusions and Clinical Relevance—CT-GFR as determined by use of a single-slice acquisition technique, low-dose of iohexol, and Patlak plot analysis correlated without bias with Inu-GFR for the right kidney and both kidneys (combined). This technique has promise as an accurate CT-GFR method that can be combined with renal morphologic evaluation.

Abstract

Objective—To compare glomerular filtration rate (GFR) as estimated from Patlak plot analysis by use of single-slice computed tomography (CT) with that obtained from clearance of plasma inulin in pigs.

Animals—8 healthy anesthetized juvenile pigs.

Procedures—All pigs underwent precontrast, whole-kidney, helical CT; postcontrast single-slice dynamic CT; and postcontrast, whole-kidney CT for volume determination. On dynamic images, corrected Hounsfield unit values were determined for each kidney and the aorta. A Patlak plot for each kidney was generated, and plasma clearance per unit volume was multiplied by renal volume to obtain whole-animal contrast clearance. Mean GFR determined via inulin clearance (Inu-GFR) was measured from each kidney and correlated to mean GFR determined via CT (CT-GFR) for the left kidney, right kidney, and both kidneys by use of linear regression and Bland-Altman analyses.

Results—CT-GFR results from 7 pigs were valid. Total and right kidney Inu-GFR were correlated with total and right kidney CT-GFR (total, R 2 = 0.85; right kidney, R 2 = 0.86). However, left kidney CT-GFR was poorly correlated with left kidney Inu-GFR (R 2 = 0.47). Bland-Altman analysis revealed no significant bias between Inu-GFR and CT-GFR for the left kidney, right kidney, or both kidneys.

Conclusions and Clinical Relevance—CT-GFR as determined by use of a single-slice acquisition technique, low-dose of iohexol, and Patlak plot analysis correlated without bias with Inu-GFR for the right kidney and both kidneys (combined). This technique has promise as an accurate CT-GFR method that can be combined with renal morphologic evaluation.

Many methods have been validated for the estimation of GFR, including renal clearance of plasma inulin, clearance of plasma creatinine, clearance of plasma iodinated contrast media, and various scintigraphic methods.1,2,a,b The classic method to determine total GFR, often described as the gold standard, is renal clearance of inulin, which is based on the determination of plasma and urinary inulin concentrations, in animals with steady-state plasma concentration of this marker. Various disadvantages are associated with this method, including the time-consuming preparation of the animal; the need for IV infusion of inulin, obtaining multiple plasma samples, and multiple urine collections with bladder or ureteral catheters; the risk of urinary tract infection; and the cumbersome chemical analysis of the concentrations of inulin in plasma and urine.1,3-5,a,b Alternatively, radioactive (isotope-labeled) markers have been used successfully because their activities in plasma and urine are easily measured in scintillation counters or can be imaged by use of a gamma camera.2,6,7

There is an increasing interest in the use of iodinated contrast medium, such as iohexol, as a GFR estimator in animals.3,4,8 Iodinated contrast medium has similar pharmacokinetics to inulin when administered IV because it does not bind to serum proteins and is completely filtered through the glomerulus, with no evidence of tubular secretion or reabsorption.3,9 Therefore, iohexol clearance should represent the GFR value. Computed tomography can precisely detect changes in the relative attenuation (density) characteristics of tissue caused by the presence of iodinated contrast medium. These attenuation changes are linearly related to tissue iodine concentration after a clinical dose of contrast medium is administered IV.10,11 On the basis of this principle, methods of GFR determination via CT have been developed in humans, pigs, and dogs, and several acquisition techniques and methods of calculation exist.12–19 Absolute CT-GFR correlates well with inulin clearance, iodine (iohexol) clearance, creatinine clearance, and scintigraphic methods of GFR determination.15,17,20 The most accurate CT-GFR methods use the Patlak plot to calculate GFR.14-17,20–22 The Patlak plot was originally developed as a method to determine the permeability of the blood-brain barrier and was first applied to determine CT-GFR in 1993.21,23 The plot is linear, and its slope represents whole-blood clearance (mL·min−1·mL of tissue).

The CT-GFR has many advantages over other methods of GFR determination, with the primary advantage being the ability to evaluate renal morphologic and functional features during the same imaging session.12-16,22 Other variables, such as renal blood flow, fractional vascular volume, and tubular dynamics, can potentially be determined during the same imaging session.17,18 The CT-GFR is rapid (requiring a few minutes), and it avoids the need for urinary catheterization or multiple samples. Contrary to scintigraphic GFR determination methods, CT-GFR can be repeated after a short delay during the same imaging session.18 Also, CT obviates the need for isolation of the patient after the procedure because of exposure to radiation.

A single-slice, low-dose iohexol Patlak plot method of CT-GFR has never been validated in anesthetized pigs. Such validation would be useful both as a research tool that could possibly replace invasive Inu-GFR determination and to extend the technique to other animal species for clinical veterinary use. The purpose of the study reported here was to validate Patlak plot CT-GFR in healthy anesthetized pigs by use of a low dose of iodinated contrast medium (iohexol) and single-slice CT acquisition by comparing CT-GFR with Inu-GFR. We hypothesized that the Patlak plot CT-GFR could provide an accurate and unbiased estimation of GFR in swine, compared with the cumbersome gold-standard procedure.

Materials and Methods

Animals—The pigs were treated according to the Canadian Council on Animal Care guidelines.24 Eight healthy young (3- to 4-month-old) male pigs were included. By random allocation, CT-GFR was acquired after Inu-GFR determination in 5 pigs and prior to Inu-GFR in 3 pigs. Three pigs were excluded and replaced because of large renal cysts seen via precontrast CT. Swine were premedicated with azaperone (2 mg·kg−1, IM) and ketamine (15 mg·kg−1, IM). After induction of anesthesia with fentanyl (0.005 mg·kg−1, IV) and propofol (4 mg·kg−1, IV [auricular vein]), the trachea was orally intubated. Anesthesia was maintained with isoflurane (1.5% to 2.5%) in O2 and continuous infusion of lidocaine (3 mg·kg−1·h−1). Lactated Ringer's solution was infused at a rate of 7 mL·kg−1·h−1 to meet maintenance needs during general anesthesia with minimally invasive surgery. Electrocardiographic leads; a pulse oximeter; and core temperature, anesthetic gases, and inspired-expired CO2 probes were placed for anesthetic monitoring. Systemic arterial pressures and heart rate were monitored continuously. Rectal temperature was continuously monitored and kept close to 38°C by use of a heating blanket.

Catheter placement—The following structures were isolated and catheterized. A 2-lumen polyurethane catheterc was placed in an external jugular vein for contrast medium administration and inulin constant infusion. A catheterd was introduced into the carotid artery for arterial blood sampling via a subcostal incision, and the left renal vein was isolated and cannulated with a catheterd for renal blood sampling. Through a ventral midline laparatomy, both ureters were isolated and cannulatede for urine collection. After instrumentation, 45 minutes was allowed to elapse to obtain a stable physiologic state prior to inulin administration.

Renal clearance of plasma inulin—The GFR was determined by means of inulin clearance. An initial bolus of inulin was administered. The bolus consisted of 2.6 g of inulinf dissolved in 10 mL of sterile water for injection and 10 mL of PBS solution and saline (0.9% NaCl) solution (up to 50 mL); pH was adjusted to 7.4 by use of NaOH. Immediately after bolus administration, a CRI of inulin was started. The rate of infusion was set at 1 mL·min−1 by use of a pump.g The CRI was prepared as follows: 0.5 g of inulin was dissolved in 2.5 mL of sterile water for injection, 6.0 mL of phosphate buffer and saline solution (up to 125 mL) was added, and pH was adjusted to 7.4 with NaOH. The total time of infusion was 110 minutes.

Approximately 30 minutes after the inulin CRI was started, after steady state was reached, the residual urine production of each kidney was collected and disposed of (time 0). Blood and urine sampling for determination of inulin concentration was begun. Starting at time = 5 minutes, venous and arterial blood was collected at 10-minute intervals initially for 3 intervals, then at 20minute intervals for 2 intervals. Alternating with this and starting at time = 10 minutes, urine was collected in separate containers for the left and right kidney, initially also at 10-minute intervals for 3 intervals and then at 20-minute intervals for 2 intervals. Total urine flow was measured at each interval.

Inulin concentration was measured in urine and plasma by use of an enzymatic assay25; Inu-GFR was considered equal to inulin clearance and was calculated by use of the standard clearance formula as follows:

article image

where CLU is the renal inulin clearance, UF is urine flow (mL·min−1), Cu is inulin urine concentration, and Ca is arterial inulin plasma concentration. Five consecutive inulin excretion rates were obtained, and the mean rate was calculated for comparison with CT-GFR results.

CT image acquisition—The CT-GFR was performed with a single-slice helical scanner.h The acquisition involved 3 steps and included baseline precontrast imaging, dynamic postcontrast single-slice imaging, and postcontrast imaging for determination of renal volume. Initial precontrast baseline imaging of both kidneys and the abdominal aorta was performed with a helical acquisition, 10-mm slice thickness, pitch of 1, and matrix of 512 X 512 (display field of view, 30 cm; scan field of view, medium, 120 kVp, and 200 mAs). By use of a pressure-injectori and associated software,j iohexol (0.25 mL·kg−1 [300 mg·mL−1; 75 mg of I·kg−1])k was injected as a bolus (rate, 4 mL·s−1). Dynamic CT acquisition was initiated simultaneously to the beginning of injection. A 10-mm-thick slice centered at the hilus of both kidneys was scanned repetitively every 4 seconds for 120 seconds. Manual breath hold was used to arrest abdominal motion, and pauses for breath were allowed at approximately 30 seconds and every 15 seconds thereafter, between slice acquisitions.

Following the dynamic acquisition, both kidneys were again scanned in their entirety for determination of renal volume. Scan variables were unchanged, except for slice thickness, which was reduced to 5 mm.

Image analysis and CT-GFR calculation—The CT-GFR determination was based on Patlak plot analysis. Regions of interest were manually drawn on an imageprocessing workstationl around the entire kidney on each dynamic slice and on the equivalent slice using the precontrast baseline acquisition, excluding the renal hilus and main vessels. Edge detection software was not used because occasional silhouetting between the renal contour and adjacent organs was present. Image window width and level were set at 150 and 20 HUs, respectively. This was repeated separately for the right and left kidneys. An ROI of constant size (22 mm2) was centered in the abdominal aorta on each dynamic and precontrast slice (Figure 1).

The mean HU value within the ROI of each kidney before administration of contrast medium was subtracted from each value after administration of contrast medium, giving a corrected kidney HU value (c[t]). The same was done for the abdominal aorta, to give a corrected aorta HU value (b[t]). Only data from the first 2 minutes after injection were used. The c(t) represents the renal tissue iodine concentration and is thus the sum of the renal blood iodine concentration and the glomerular filtrate iodine concentration. The b(t) represents aortic blood iodine concentration, which was assumed to be proportional to renal blood iodine concentration. A Patlak plot for each kidney was then generated by plotting ∫b(t) dt/b(t) against c(t)/b(t). A TAC was generated for the aorta to calculate ∫b(t) dt, which is the area under the aortic TAC. This area represents the shape that an organ with 100% extraction efficiency would have and has units of HU X time (s). The ∫b(t) dt/b(t) is this area divided by the aortic blood iodine concentration; it has units of time and is also known as normalized time. The c(t) differs from ∫b(t) dt because the renal ROI from which it is calculated does not necessarily have 100% extraction efficiency (ie, within the ROI volume, there is nonfunctional interstitial tissue and there may be a number of nonfunctioning glomeruli). The c(t)/b(t) is unitless: the units of tissue concentration of iodine (represented by the HU value) cancel.

Figure 1—
Figure 1—

Dynamic CT images centered at the renal hilus obtained at maximal aortic enhancement with iohexol in a pig. White arrow indicates the aorta, the white asterisk indicates the left kidney, and the black asterisk indicates the right kidney. Dorsal is at the top of the image and left is to the left of the image. Notice ROIs placed around each kidney in the image on the right.

Citation: American Journal of Veterinary Research 68, 3; 10.2460/ajvr.68.3.297

Figure 2—
Figure 2—

Time attenuation curves related to the concentration of iohexol in the aortic blood and right and left kidneys of a pig. The gap at 36 seconds was caused by a breath-hold misregistration error.

Citation: American Journal of Veterinary Research 68, 3; 10.2460/ajvr.68.3.297

The resulting Patlak plot is linear. The slope of °b(t) dt/b(t) versus c(t)/b(t) is the blood clearance per unit volume (mL·min−1·mL of tissue) and is also known as α/V. If the filtration fraction was 100%, the slope of ∫b(t)dt/b(t) versus c(t)/b(t) would be unity. This slope represents whole-blood clearance and must be corrected with PCV to obtain plasma clearance GFR (slope X 1 − PCV). It is generally accepted that a Patlak plot where the correlation between ∫b(t) dt/b(t) and c(t)/b(t) is ≥ 0.95 is reliable.17

Several assumptions are inherent to the Patlak model: it is a 2-compartment model that assumes that the net transfer of solute (iohexol) across a barrier (glomerular membrane) is unidirectional and that once transferred, all of this solute remains in the second compartment (the kidney).23 The model also assumes that homogeneous mixing of the solute (iohexol) in plasma has occurred.17,19

Renal volume was calculated by measuring the cross-sectional area of an ROI manually drawn on each kidney on each slice. Image window width and level were set at 80 and 25 HUs, respectively, for ROI determination. The total kidney cross-sectional area was obtained by summing these results. This was then multiplied by slice thickness to give the volume. Each kidney volume was calculated separately. Plasma clearance per unit volume for each kidney (mL·min−1·mL of tissue) was then multiplied by renal volume to obtain global contrast clearance (mL·min−1). Data are provided uncorrected for the animal's total weight.

Statistical analysis—Statistical analysis was performed with statistical software.m The mean and SD of CT-GFR and Inu-GFR were obtained. Because of the small sample size, Wilcoxon signed rank tests were used to compare the CT-GFR values for the right kidney versus left kidney as well as Inu-GFR values for the right kidney versus left kidney. Then, linear regression was applied to compare CT-GFR with Inu-GFR for the right kidney, left kidney, and the sum of both kidneys (total GFR). Agreement was evaluated by use of Bland-Altman analysis. The Bland-Altman method is used to plot the difference between 2 diagnostic tests (eg, CT-GFR and Inu-GFR) against the mean of results of both tests together.26 This technique estimates the degree to which 1 test agrees with another, can detect whether a consistent over- or underestimation is present, and determines the upper and lower values for the difference of agreement. When 2 tests are in perfect agreement, points in the Bland-Altman plot will be clustered around a line representing 0 on the y-axis. These tests can then be used interchangeably in a clinical situation. Significance was set at P < 0.05 for all tests.

Figure 3—
Figure 3—

Patlak plot for a pig with regression equations for the right (RK) and left (LK) kidneys.

Citation: American Journal of Veterinary Research 68, 3; 10.2460/ajvr.68.3.297

Results

The pigs ranged in weight from 17.5 to 37.0 kg (mean 26.5 kg), and PCV ranged from 26% to 31% (mean, 28%). The Patlak plot for 1 pig had large variability (right kidney Patlak, R2 = 0.64; left kidney Patlak, R2 = 0.27). With such high variability, CT-GFR results for this pig were considered invalid and results from this pig were eliminated from all statistical analysis and results.

Peak aortic enhancement occurred at 4 seconds (n = 2 pigs) or 8 seconds (5) after injection of contrast medium. The aortic time attenuation curves revealed an initial high enhancement peak, with a second much lower peak at 20 to 35 seconds. Progressively smaller peaks occurred approximately every 20 seconds thereafter (Figure 2).

Initial peak renal parenchymal enhancement occurred at 8 to 20 seconds (mean, 14 seconds) and thus occurred 4 to 8 seconds after maximal aortic enhancement. In 5 pigs, both kidneys were enhanced maximally simultaneously. In the remaining 2 pigs, a 4-second delay occurred between the kidneys; in 1 pig, the right kidney was enhanced first, and in the other, the left kidney was enhanced first. After this initial enhancement peak, much smaller peaks occurred approximately every 20 seconds. Despite these small undulations in the renal TAC, an overall gradual increase in renal parenchymal HU values followed the initial enhancement peak until the end of data acquisition at 120 seconds. In 4 pigs, this overall increase in renal parenchymal attenuation was equivalent to or surpassed the HU value of the initial enhancement peak (Table 1).

Table 1—

Range and mean (n = 7 pigs) values of aortic and renal attenuation values obtained via CT before (precontrast) and after (postcontrast) administration of iohexol.

VariableRange(HU)Mean(HU)
Precontrast aortic attenuation36.7–41.938.6
Precontrast renal attenuation (both kidneys)24.8–32.930.4
Postcontrast aortic peak attenuation (uncorrected)152.9–233.4184.5
Postcontrast renal peak attenuation (uncorrected)43.6–56.651.4
Figure 4—
Figure 4—

Linear regression graphs of Inu-GFR versus CT-GFR (mL·min−1) for the LK (A), RK (B), and total of both kidneys (C) in 7 pigs. Straight line represents the linear regression between CT-GFR and Inu-GFR for the LK, RK, and total of both kidneys.

Citation: American Journal of Veterinary Research 68, 3; 10.2460/ajvr.68.3.297

Figure 5—
Figure 5—

Bland-Altman graphs of Inu-GFR and CT-GFR for the LK (A), RK (B), and total of both (total) kidneys (C) in 7 pigs. Straight line represents the Bland-Altman relationship between the difference between CT-GFR and Inu-GFR (y-axis) and the mean of CT-GFR and Inu-GFR.

Citation: American Journal of Veterinary Research 68, 3; 10.2460/ajvr.68.3.297

In 6 pigs, contrast medium was seen in the collecting system of both kidneys by the time the renal volume scan was obtained (6 to 7 minutes after injection). In 1 pig, contrast medium was not seen in the collecting system. Moderate hydronephrosis was present in the 2 first-tested pigs, probably as a result of Foley catheter inflation. Foley catheters were left uninflated for the remaining pigs; no hydronephrosis occurred in those pigs.

For all pigs, outlier values were seen in the Patlak plots at points corresponding with initial aortic or renal attenuation peaks (Figure 3). These outlier points were removed for calculation of the Patlak regression equations.

The CT-GFR and Inu-GFR data distribution for the left kidney, right kidney, and total kidneys was normal (Table 2). One pig had lower GFR than all of the others, particularly when evaluated via CT.

Table 2—

Results of determination of GFR (mL•min−1) in 7 pigs as a function of kidney side and testing method.

KidneyMethodMean(SD)Median
LeftInu-GFR40.9(12.5)41.2
CT-GFR34.3(14.0)35.6
RightInu-GFR41.0(18.1)38.6
CT-GFR37.1 (16.0)36.9
TotalInu-GFR82.0(28.6)80.3
CT-GFR71.4(29.5)72.6

No significant (P = 0.938) differences were found between the left kidney and right kidney Inu-GFR. Although not significant (P = 0.156), left kidney function was often slightly lower than right kidney function as measured via CT.

For linear regression, Inu-GFR (gold standard) was designated as the independent variable and CT-GFR as the dependent variable (Table 3; Figure 4). For the left kidney, CT-GFR was a poor predictor of Inu-GFR, with R2 of 0.47 and a slope that was not significantly different from 0 (P = 0.09). Right kidney Inu-GFR was more predictable (R2 = 0.86), with a slope that was significantly (P = 0.002) different from 0 and not significantly different from 1 (P = 0.2). Total Inu-GFR was almost as closely predicted (R2 = 0.85), and the slope was significantly (P = 0.003) different from 0 and not significantly different from 1 (P = 0.7).

Table 3—

Results of linear regression analysis of the relationship between Inu-GFR and CT-GFR for the left kidney, right kidney, and total CT-GFR in 7 pigs.

KidneyR2Slope (SE)P valueIntercept (SE)P value
Left0.470.77 (0.36)0.0902.98(15.48)0.855
Right0.860.82(0.15)0.0023.43 (6.45)0.618
Total0.850.94(0.18)0.003−6.43(15.28)0.691

The mean difference between Inu-GFR and CT-GFR was 6.63 mL·min−1 (95% CI, ± 20.8 mL·min−1) for the left kidney, 3.9 mL·min−1 (95% CI, ± 13.2mL·min−1) for the right kidney, and 10.54 mL·min−1 (95% CI, ± 22.4 mL·min−1) for the total. Although CT-GFR often underestimated Inu-GFR (Figure 5), there was no significant relationship (bias) between the difference in Inu-GFR and CT-GFR and the mean of these 2 measurements for the left kidney (P = 0.75), right kidney (P = 0.49), and total (P = 0.87).

Linear regression statistics were repeated with 1 pig eliminated from the data set because its GFR was below the reference limit value (Table 4). Good prediction of right kidney Inu-GFR (R2 = 0.83) and total Inu-GFR (R2 = 0.81) was still evident via CT-GFR. Although the slopes of both of the regression equations were farther from 1.0, they remained significantly different from 0. The intercepts were still not significantly different from 0. Poor prediction of left kidney Inu-GFR by CT-GFR remained (R2 = 0.21). Bland-Altman analysis revealed no significant bias for the left kidney (P = 0.61), right kidney (P = 0.22), and total (P = 0.40).

Table 4—

Results of linear regression analysis of the relationship between Inu-GFR and CT-GFR for the left kidney, right kidney, and total CT-GFR in 6 pigs.*

KidneyR2RSlope (SE)P valueIntercept (SE)P value
Left0.210.460.36(0.34)0.36122.97(15.60)0.215
Right0.830.910.68(0.15)0.01110.93(7.19)0.204
Total0.810.900.74(0.18)0.01414.94(16.08)0.691

One pig from the original group of 7 was removed from the study because of low CT-GFR results.

Discussion

For the present study, a single-slice CT acquisition technique, Patlak plot analysis, and a low IV dose of contrast medium were selected. The single-slice technique is limited in that it requires extrapolation of GFR information from a single slice to that of the entire kidney.17,20,21 Therefore, in diseased kidneys in which there is a large variability in function, the single-slice technique may be inaccurate. However, the data generated from the single-slice acquisition contain a greater number of time points and thus allows generation of more complete Patlak plots and TACs. In addition, this acquisition technique allows other mathematical models to be applied to the data and for separate data to be acquired from the cortex and medulla.

Patlak plot analysis is a 2-compartment model based on the assumption that the GFR is equivalent to the influx constant (Ki) of a solute (ie, contrast medium) across a barrier (ie, the glomerular membrane). The model may be applied when solute flow is unidirectional, which is true when evaluating GFR because back diffusion across the glomerulus does not occur. Linearity of the Patlak plot is related to the fact that solute flow is unidirectional.23 The entire amount of diffused solute must be present within the organ parenchyma at the time of sampling, making it important to perform the analysis prior to the presence of contrast medium within the ureters. For this reason, our analysis included only the data obtained up to 120 seconds. Too short of an acquisition period (10 to 15 seconds) results in inaccurate CT-GFR determination.15 Several other conditions must also be met.23 The Patlak model accounts for the fact that there may be a change in time of the test solute in plasma, which is clearly true with a bolus of contrast medium within the aorta.23 A limitation in the model is that it does not account for interstitial space as a third compartment; therefore, CT-GFR in patients with increased interstitial space may be inaccurate.14

The low dose of contrast medium (0.25 mg·k−1 [75 mg of I·kg−1]) allowed calculation of GFR while limiting the risk of CMIN. This dose was chosen on the basis of another CT-GFR study.20 The CMIN has become a major concern in human medicine, in which frequency rates of severe CMIN can be as high as 18%.27 Contrast medium dose is recognized to be one of the major contributing factors for CMIN in humans, especially when doses as high as 900 to 1,800 mg of I/kg are used.27

The shape of the renal time attenuation curves was similar among all individuals in the study reported here and similar to that seen with camera-based radiopharmaceutical plasma clearance methods2 and 2 other CT-GFR studies.19,21 It is likely that the time to maximal enhancement, maximal HU value, and shape of the renal TAC would be altered in animals with decreased renal function and that the alteration of these variables could be related to specific disease processes. In this study, for 1 pig that had the lowest CT-GFR function, the renal TACs were flattest with the longest time to maximal enhancement (20 seconds) and the lowest corrected peak enhancement. More complete renal function curves could be constructed with a longer data acquisition period (ie, 5 to 10 minutes).

A recirculation effect was seen in the renal and aortic TACs and consisted of a series of small undulations after the initial maximal enhancement peak. This indicated that iohexol was not homogeneously mixed in blood after injection. Because the Patlak model assumes that homogeneous mixing of iohexol in plasma has occurred, this recirculation effect has been mentioned as proof that the model may not be strictly applicable to GFR within the first few seconds after bolus injection.11 For this reason, it has been suggested that Patlak plot analysis should be delayed until after 2 or 3 circulation peaks.11 The outlier data points seen in Patlak plots of pigs of the present study seemed to represent these early recirculation peaks and were consequently removed from the Patlak plot. In doing this, Patlak plot analysis performed as of time 0 in the study group was successful in this study as well as in another study.17

As expected, no difference in GFR was found between the left and right kidneys by use of either inulin plasma clearance or CT. This was also consistent with CT-GFR studies evaluating relative renal function.28 However, left kidney CT-GFR was a poor predictor of left kidney Inu-GFR. No clear explanation for this discrepancy was found, but this relationship indicated a larger and more variable discrepancy between left kidney Inu-GFR and CT-GFR. Renal function, image analysis, image acquisition, and sample analysis factors were identical for both kidneys. A similar discrepancy between right renal CT-GFR and right renal scintigraphic GFR has been explained by temporal differences in attenuation of the kidneys.28 In pigs, the left kidney is generally slightly more caudally located, which could result in a slight delay in contrast medium arriving at this kidney and a subsequent delay in the time attenuation curve for this kidney. This slight delay would affect the CT-GFR results but not the Inu-GFR results. However, no consistent delay in the left renal TAC was observed in our pigs. Another explanation is that extrapolation of CT-GFR from a single slice of kidney led to a consistently greater degree of error for the left kidney. The percentage of nonfunctioning renal tissue may have been overrepresented in the slices acquired in the left kidneys, which led to significantly lower left kidney CT-GFR values.

Right kidney CT-GFR predicted right kidney Inu-GFR accurately and without observable bias in this small population of swine. The CT-GFR of only the right kidney may therefore be a sufficient estimate of GFR when CT-GFR is used as a research tool to evaluate normal renal function. This would reduce the amount of necessary image analysis by half. Total CT-GFR also predicted total Inu-GFR accurately and without bias. This strong positive correlation detected between the right kidney and total kidneys was comparable to other studies17,20 that used the same CT-GFR analysis method in which total CT-GFR had a correlation of R = 0.87 to 0.92 with blood clearance of creatinine in humans with and without diabetes. An underestimation of GFR by use of CT has been reported by others.17,19 Transient CMIN, vascular effects related to general anesthesia, and the Faraeus effect have all been used to explain the lower GFR value obtained by use of CT.10,17,19 Underestimation of renal volume would also lead to an underestimation with CT-GFR. In patients with increased interstitial space, CT-GFR overestimates the value obtained by use of the test standard.14 Although no such over or underestimation was detected in the present study via Bland-Altman analysis, it is possible that had a larger test population been used, a significant bias between CT-GFR and Inu-GFR would have been found.

One pig was eliminated from the study because of invalidation of the Patlak plot analysis and CT-GFR results. A breath-hold misregistration error, resulting in excessive patient motion, occurred at 20 seconds after injection. This caused an early gap in the aortic and renal TACs and likely caused a calculation error within the Patlak plot analysis. Because of the 2-minute image acquisition period, 1 or more breath-hold misregistration errors occurred in every pig in this study; however, most were late in the image acquisition and had no apparent effect on the results.

Total GFR as determined via Inu-GFR and technetium 99mTc-pentetate plasma clearance in healthy juvenile anesthetized pigs ranges from 40 to 100 mL·min−1 for pigs with a similar weight as our study population.18,29,30 The Inu-GFR and CT-GFR results were within this range for all pigs in this study, except 1, in which Inu-GFR was near the lower limit of the range at 42.6 mL·min−1 and CT-GFR was low at 19.5 mL·min−1. Because of this finding, it was unclear whether or not this pig should be excluded; thus, results were presented both with and without this pig. The discrepancy in function between Inu-GFR and CT-GFR in this pig may indicate that in individuals with renal function near the lower limit of the range in healthy animals, CT-GFR may underestimate renal function. Another explanation is a temporal discrepancy between the CT-GFR results and the Inu-GFR results, which were obtained at least 1 hour from the CT-GFR results and over a longer period. A temporal reduction in GFR during the CT acquisition could have been the result of a transient reduction in renal blood flow or glomerular function secondary to contrast medium vascular effects, hypotension, hypovolemia, or transient excessive anesthetic depth.

The CT-GFR performed with a single-slice acquisition technique, low dose of iohexol, and Patlak plot analysis correlated highly and without bias with inulin plasma clearance GFR for the right kidney and total kidneys. This technique has promising clinical relevance in the development of an accurate CT-GFR method that can be combined with renal morphologic evaluation. Further studies on a larger number of individuals, in other species, and in animals with abnormal renal function are needed to investigate the clinical usefulness of this technique.

ABBREVIATIONS

GFR

Glomerular filtration rate

CT

Computed tomography

CT-GFR

GFR determined via CT

Inu-GFR

GFR determined via inulin clearance

CRI

Constant rate infusion

ROI

Region of interest

HU

Hounsfield unit

TAC

Time-attenuation curve

CMIN

Contrast medium–induced nephrotoxicity

a.

Barthez P. Validation de méthodes de mesure isotopique de l'hémodynamique et de la fonction rénale chez l'animal. PhD thesis, Department of Small Animals, Ecole Nationale Vétérinaire de Lyon, France: Université Claude-Bernard, 2000;172.

b.

Frennby B. Use of iohexol clearance to determine the glomerular filtration rate. A comparison between different clearance techniques in man and animal. PhD thesis, Department of diagnostic radiology, University of Lund, Malmö University Hospital, S-20502, Malmö, Sweden: 1996;63.

c.

CS-16702, Arrow International, Reading, Pa.

d.

MRE-065, Braintree Scientific, Braintree, Mass.

e.

V-PFC6-55, Cook Veterinary Products, Bloomington, Ind.

f.

I3754, Sigma-Aldrich Canada Ltd, Oakville, ON, Canada.

g.

55-2222, Harvard Apparatus, Holliston, Mass.

h.

Hi-speed ZXi, General Electric, Milwaukee, Wis.

i.

402 syringe pump, Gilson, Middleton, Wis.

j.

Gilson Serial Input Output Channel (GSIOC), Gilson, Middleton, Wis.

k.

Omnipaque 300, Amersham Health Inc, Oakville, ON, Canada.

l.

Advantage (AW), version 4.0, General Electric, Milwaukee, Wis.

m.

SAS, version 8.02, SAS Institute Inc, Cary, NC.

References

  • 1

    Levey A, Madaio M, Perrone R. Laboratory assessment of renal disease: clearance, urinanalysis, and renal biopsy. In:Brenner BM, Rector FC, ed.The kidney. 4th ed.Philadelphia: WB Saunders Co, 1991;919925.

    • Search Google Scholar
    • Export Citation
  • 2

    Daniel GB, Mitchell SK, Mawby D, et al. Renal nuclear medicine: a review. Vet Radiol Ultrasound 1999;40:572587.

  • 3

    Brown SA, Finco DR, Boudinot FD, et al. Evaluation of a single injection method, using iohexol, for estimating glomerular filtration rate in cats and dogs. Am J Vet Res 1996;57:105110.

    • Search Google Scholar
    • Export Citation
  • 4

    Miyamoto K. Use of plasma clearance of iohexol for estimating glomerular filtration rate in cats. Am J Vet Res 2001;62:572575.

  • 5

    Rogers KS, Komkov A, Brown SA, et al. Comparison of four methods of estimating glomerular filtration rate in cats. Am J Vet Res 1991;52:961964.

    • Search Google Scholar
    • Export Citation
  • 6

    Barthez PY, Hornof WJ, Cowgill LD, et al. Comparison between the scintigraphic uptake and plasma clearance of 99mTc-diethylenetriaminepentacetic acid (DTPA) for the evaluation of the glomerular filtration rate in dogs. Vet Radiol Ultrasound 1998;39:470474.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 7

    Uribe D, Krawiec DR, Twardock AR, et al. Quantitative renal scintigraphic determination of the glomerular filtration rate in cats with normal and abnormal kidney function, using 99mTc-diethylenetriaminepentaacetic acid. Am J Vet Res 1992;53:11011107.

    • Search Google Scholar
    • Export Citation
  • 8

    Moe L, Heiene R. Estimation of glomerular filtration rate in dogs with 99M-Tc-DTPA and iohexol. Res Vet Sci 1995;58:138143.

  • 9

    Morris T, Fischer H. The pharmacology of intravascular radiocontrast media. Ann Rev Pharmacol Toxicol 1986;26:143160.

  • 10

    Lerman LO, Rodriguez-Porcel M, Romero JC. The development of x-ray imaging to study renal function. Kidney Int 1999;55:400416.

  • 11

    Blomley MJ, Dawson P. Review article: the quantification of renal function with enhanced computed tomography. Br J Radiol 1996;69:989995.

  • 12

    el-Diasty TA, Shokeir AA, el-Ghar ME, et al. Contrast enhanced spiral computerized tomography in live kidney donors: a single session for anatomical and functional assessment. J Urol 2004;171:3134.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 13

    El-Ghar ME, Shokeir AA, El-Diasty TA, et al. Contrast enhanced spiral computerized tomography in patients with chronic obstructive uropathy and normal serum creatinine: a single session for anatomical and functional assessment. J Urol 2004;172:985988.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 14

    Hackstein N, Bauer J, Hauck EW, et al. Measuring single-kidney glomerular filtration rate on single-detector helical CT using a two-point Patlak plot technique in patients with increased interstitial space. AJR Am J Roentgenol 2003;181:147156.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 15

    Hackstein N, Cengiz H, Rau WS. Contrast media clearance in a single kidney measured on multiphasic helical CT: results in 50 patients without acute renal disorder. AJR Am J Roentgenol 2002;178:111118.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 16

    Hackstein N, Puille MF, Bak BH, et al. Measurement of single kidney contrast media clearance by multiphasic spiral computed tomography: preliminary results. Eur J Radiol 2001;39:201208.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 17

    Tsushima Y, Blomley MJ, Kusano S, et al. Use of contrast-enhanced computed tomography to measure clearance per unit renal volume: a novel measurement of renal function and fractional vascular volume. Am J Kidney Dis 1999;33:754760.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 18

    Krier JD, Ritman EL, Bajzer Z, et al. Noninvasive measurement of concurrent single-kidney perfusion, glomerular filtration, and tubular function. Am J Physiol Renal Physiol 2001;281: F630F638.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 19

    O'Dell-Anderson KJ, Twardock R, Grimm JB, et al. Determination of glomerular filtration rate in dogs using contrast-enhanced computed tomography. Vet Radiol Ultrasound 2006;47:127135.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 20

    Tsushima Y, Blomley MJ, Okabe K, et al. Determination of glomerular filtration rate per unit renal volume using computerized tomography: correlation with conventional measures of total and divided renal function. J Urol 2001;165:382385.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 21

    Dawson P, Peters M. Dynamic contrast bolus computed tomography for the assessment of renal function. Invest Radiol 1993;28:10391042.

  • 22

    Hackstein N, Wiegand C, Rau WS, et al. Glomerular filtration rate measured by using triphasic helical CT with a two-point Patlak plot technique. Radiology 2004;230:221226.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 23

    Patlak CS, Blasberg RG, Fenstermacher JD. Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data. J Cereb Blood Flow Metab 1983;3:17.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 24

    Olfert E, Cross B, McWilliam A. Manuel sur le soin et l'utilisation des animaux d'expérimentation. 2nd ed.Ottawa: Conseil Canadien de Protection des Animaux, 1993.

    • Search Google Scholar
    • Export Citation
  • 25

    Kuehnle HF, von Dahl K, Schmidt FH. Fully enzymatic inulin determination in small volume samples without deproteinization. Nephron 1992;62:104107.

  • 26

    Bland J, Altman D. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307310.

  • 27

    Nyman U, Almen T, Aspelin P, et al. Contrast-medium-induced nephropathy correlated to the ratio between dose in gram iodine and estimated GFR in ml/min. Acta Radiol 2005;46:830842.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 28

    Nilsson H, Wadstrom J, Andersson LG, et al. Measuring split renal function in renal donors: can computed tomography replace renography? Acta Radiol 2004;45:474480.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 29

    Troncy E, Francoeur M, Salazkin I, et al. Extra-pulmonary effects of inhaled nitric oxide in swine with and without phenylephrine. Br J Anaesth 1997;79:631640.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 30

    Eskild-Jensen A, Rehling M, Nielsen AS, et al. Use of plasma clearance of technetium Tc 99m pentetate to estimate renal clearance during postnatal development in pigs. Am J Vet Res 2002;63:12031206.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Figure 1—

    Dynamic CT images centered at the renal hilus obtained at maximal aortic enhancement with iohexol in a pig. White arrow indicates the aorta, the white asterisk indicates the left kidney, and the black asterisk indicates the right kidney. Dorsal is at the top of the image and left is to the left of the image. Notice ROIs placed around each kidney in the image on the right.

  • Figure 2—

    Time attenuation curves related to the concentration of iohexol in the aortic blood and right and left kidneys of a pig. The gap at 36 seconds was caused by a breath-hold misregistration error.

  • Figure 3—

    Patlak plot for a pig with regression equations for the right (RK) and left (LK) kidneys.

  • Figure 4—

    Linear regression graphs of Inu-GFR versus CT-GFR (mL·min−1) for the LK (A), RK (B), and total of both kidneys (C) in 7 pigs. Straight line represents the linear regression between CT-GFR and Inu-GFR for the LK, RK, and total of both kidneys.

  • Figure 5—

    Bland-Altman graphs of Inu-GFR and CT-GFR for the LK (A), RK (B), and total of both (total) kidneys (C) in 7 pigs. Straight line represents the Bland-Altman relationship between the difference between CT-GFR and Inu-GFR (y-axis) and the mean of CT-GFR and Inu-GFR.

  • 1

    Levey A, Madaio M, Perrone R. Laboratory assessment of renal disease: clearance, urinanalysis, and renal biopsy. In:Brenner BM, Rector FC, ed.The kidney. 4th ed.Philadelphia: WB Saunders Co, 1991;919925.

    • Search Google Scholar
    • Export Citation
  • 2

    Daniel GB, Mitchell SK, Mawby D, et al. Renal nuclear medicine: a review. Vet Radiol Ultrasound 1999;40:572587.

  • 3

    Brown SA, Finco DR, Boudinot FD, et al. Evaluation of a single injection method, using iohexol, for estimating glomerular filtration rate in cats and dogs. Am J Vet Res 1996;57:105110.

    • Search Google Scholar
    • Export Citation
  • 4

    Miyamoto K. Use of plasma clearance of iohexol for estimating glomerular filtration rate in cats. Am J Vet Res 2001;62:572575.

  • 5

    Rogers KS, Komkov A, Brown SA, et al. Comparison of four methods of estimating glomerular filtration rate in cats. Am J Vet Res 1991;52:961964.

    • Search Google Scholar
    • Export Citation
  • 6

    Barthez PY, Hornof WJ, Cowgill LD, et al. Comparison between the scintigraphic uptake and plasma clearance of 99mTc-diethylenetriaminepentacetic acid (DTPA) for the evaluation of the glomerular filtration rate in dogs. Vet Radiol Ultrasound 1998;39:470474.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 7

    Uribe D, Krawiec DR, Twardock AR, et al. Quantitative renal scintigraphic determination of the glomerular filtration rate in cats with normal and abnormal kidney function, using 99mTc-diethylenetriaminepentaacetic acid. Am J Vet Res 1992;53:11011107.

    • Search Google Scholar
    • Export Citation
  • 8

    Moe L, Heiene R. Estimation of glomerular filtration rate in dogs with 99M-Tc-DTPA and iohexol. Res Vet Sci 1995;58:138143.

  • 9

    Morris T, Fischer H. The pharmacology of intravascular radiocontrast media. Ann Rev Pharmacol Toxicol 1986;26:143160.

  • 10

    Lerman LO, Rodriguez-Porcel M, Romero JC. The development of x-ray imaging to study renal function. Kidney Int 1999;55:400416.

  • 11

    Blomley MJ, Dawson P. Review article: the quantification of renal function with enhanced computed tomography. Br J Radiol 1996;69:989995.

  • 12

    el-Diasty TA, Shokeir AA, el-Ghar ME, et al. Contrast enhanced spiral computerized tomography in live kidney donors: a single session for anatomical and functional assessment. J Urol 2004;171:3134.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 13

    El-Ghar ME, Shokeir AA, El-Diasty TA, et al. Contrast enhanced spiral computerized tomography in patients with chronic obstructive uropathy and normal serum creatinine: a single session for anatomical and functional assessment. J Urol 2004;172:985988.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 14

    Hackstein N, Bauer J, Hauck EW, et al. Measuring single-kidney glomerular filtration rate on single-detector helical CT using a two-point Patlak plot technique in patients with increased interstitial space. AJR Am J Roentgenol 2003;181:147156.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 15

    Hackstein N, Cengiz H, Rau WS. Contrast media clearance in a single kidney measured on multiphasic helical CT: results in 50 patients without acute renal disorder. AJR Am J Roentgenol 2002;178:111118.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 16

    Hackstein N, Puille MF, Bak BH, et al. Measurement of single kidney contrast media clearance by multiphasic spiral computed tomography: preliminary results. Eur J Radiol 2001;39:201208.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 17

    Tsushima Y, Blomley MJ, Kusano S, et al. Use of contrast-enhanced computed tomography to measure clearance per unit renal volume: a novel measurement of renal function and fractional vascular volume. Am J Kidney Dis 1999;33:754760.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 18

    Krier JD, Ritman EL, Bajzer Z, et al. Noninvasive measurement of concurrent single-kidney perfusion, glomerular filtration, and tubular function. Am J Physiol Renal Physiol 2001;281: F630F638.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 19

    O'Dell-Anderson KJ, Twardock R, Grimm JB, et al. Determination of glomerular filtration rate in dogs using contrast-enhanced computed tomography. Vet Radiol Ultrasound 2006;47:127135.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 20

    Tsushima Y, Blomley MJ, Okabe K, et al. Determination of glomerular filtration rate per unit renal volume using computerized tomography: correlation with conventional measures of total and divided renal function. J Urol 2001;165:382385.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 21

    Dawson P, Peters M. Dynamic contrast bolus computed tomography for the assessment of renal function. Invest Radiol 1993;28:10391042.

  • 22

    Hackstein N, Wiegand C, Rau WS, et al. Glomerular filtration rate measured by using triphasic helical CT with a two-point Patlak plot technique. Radiology 2004;230:221226.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 23

    Patlak CS, Blasberg RG, Fenstermacher JD. Graphical evaluation of blood-to-brain transfer constants from multiple-time uptake data. J Cereb Blood Flow Metab 1983;3:17.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 24

    Olfert E, Cross B, McWilliam A. Manuel sur le soin et l'utilisation des animaux d'expérimentation. 2nd ed.Ottawa: Conseil Canadien de Protection des Animaux, 1993.

    • Search Google Scholar
    • Export Citation
  • 25

    Kuehnle HF, von Dahl K, Schmidt FH. Fully enzymatic inulin determination in small volume samples without deproteinization. Nephron 1992;62:104107.

  • 26

    Bland J, Altman D. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307310.

  • 27

    Nyman U, Almen T, Aspelin P, et al. Contrast-medium-induced nephropathy correlated to the ratio between dose in gram iodine and estimated GFR in ml/min. Acta Radiol 2005;46:830842.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 28

    Nilsson H, Wadstrom J, Andersson LG, et al. Measuring split renal function in renal donors: can computed tomography replace renography? Acta Radiol 2004;45:474480.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 29

    Troncy E, Francoeur M, Salazkin I, et al. Extra-pulmonary effects of inhaled nitric oxide in swine with and without phenylephrine. Br J Anaesth 1997;79:631640.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 30

    Eskild-Jensen A, Rehling M, Nielsen AS, et al. Use of plasma clearance of technetium Tc 99m pentetate to estimate renal clearance during postnatal development in pigs. Am J Vet Res 2002;63:12031206.

    • Crossref
    • Search Google Scholar
    • Export Citation

Advertisement