565

Glomerular Filtration Rate:

Estimation from Fractional Renal
Accumulation of 99mTc-DTPA
(Stan nous)

aamTc diethylenetriaminepentaacetic acid (stannous) is cleared from the blood by

Gary F. Gate&
glomerular filtration. However, calculation of the glomerular filtration rate has required
analysis of serial blood samples drawn over intervals of several hours in order to
determine the radionuclide clearance rate. This lengthy procedure can be circumvented
by the direct scintigraphic determination of fractional radionuclide accumulation within
each kidney, occurring at a specified time interval after radionuclide administration.
This value has been correlated with 24 hr creatinine clearance values in a group of
hospitalized patients. From these data, a formula was derived allowing for estimation
of total glomerular filtration rate, as well as individually for each kidney, without
requiring blood samples. This procedure needs only 6 mm of patient time.

The determination of the glomerular filtration rate (GFR) as measured by

creatinine clearance has been a recognized means of assessing renal function
for many decades [1 ]. During the past few years, several accurate radionuclide
techniques have been developed in order to determine GFR, both as a total
integral value of renal function and individually for each kidney. However, these
procedures require that serial blood samples be obtained in order to perform the
computations necessary to calculate GFR. Thus the time interval needed for
acquiring the multiple blood samples, which may be up to 4 hr, combined with
the sample counting and subsequent back extrapolation of data in order to
determine the radionuclide clearance rate impose logistical constraints upon a
busy nuclear medicine department.
This report describes a rapid, accurate method of determining GFR, both as a
total value and individually for each kidney, by use of 99mTcDTPA (diethylene-
tniaminepentaacetic acid). Blood sampling techniques are not necessary, and the
entire study requires only 6 mm of patient time. The validity has been confirmed
by comparing the radionuclide-computed GFR values with 24 hr creatinine
clearance values in 31 hospitalized patients. The reproducibility of results has
also been determined by performing repeated studies in the same patients and
comparing the results.

Subjects and Methods

Received June 5, 1981 ; accepted after revision
Thirty-one hospitalized patients including two with renal transplantations were studied
October 30, 1 981.
35 times (one patient four times, another twice). Twenty-four hour creatinine clearance
Presented at the annual meeting of the Amen- values were determined in these individuals and used for data analysis. The average age
can Roentgen Ray Society, San Francisco, March
of these patients was 57.3 years (range, 1 7-86): 1 5 were men and 1 6 were women.
1981.
S9mTcDTpA (Sn complexed, Diagnostic Isotopes) 3 mCi [1 1 i MBq] was the radionuclide
I Department of Nuclear
Medicine, Good Sa-
maritan Hospital and Medical Center, 1015 NW.
used in all patients. Thin-layer chromatography of the radiopharmaceutical preparation
22nd Ave., Portland, OR 97210. showed over 99.9#{176}/o binding of aamTc to DTPA. A 1 mm preinjection count of the administered
dose was performed by placing the syringe 30 cm from the center of the parallel-hole
AJR 138:565-570, March 1982
0361 -803X/82/ 1383-0565 $00.00 medium-energy collimator of a gamma camera which, in turn, was interfaced with a digital
© American Roentgen Ray Society computer. After the conclusion of the patient part of the study, a similar 1 mm postinjection
566 GATES AJR:138, March 1982

syringe count was performed. The radionuclide was rapidly injected

intravenously into the patients who were prepositioned posteriorly
before the gamma
camera, and the subsequent study was stored
on a 1 28 x in 1 5 sec intervals.
1 28 matrix The same matrix size
was used for the 1 mm syringe count data. Data acquisition was
initiated at the moment of injection.
After the conclusion of the 6 mm examination, a composite image
of the entire study was created and processed using 50% back-
ground subtraction. A light pen was used to outline each kidney
with its separate background region as an area of interest. The
preinjection and postinjection static 1 mm syringe counts were
determined and, by subtraction, the net administered counts were A
calculated. Each patient’s height and weight were measured (in
centimeters and kilograms) at the time of the examination.
Several variables were tested during this study in order to deter-
mine the method by which the radionuclide GFA calculation would
have the closest correlation with the 24 hr creatinine clearance
value. Two penirenal areas were used for background correction:
(1 ) a semilunar region inferior and somewhat lateral to the kidney
trying to avoid the spleen, liver, and iliac vessels and (2) a circum-
ferential ring that surrounded the entire kidney (fig. 1 ). Net radio-
nuclide counts within the kidneys were determined by subtraction
of the activity in the background region (whose area was normalized
in size to its respective renal area) from gross renal counts. This
calculation was performed at various time intervals after tracer
C D
arrival in the kidneys: 0-1 mm, 1-2 mm, 2-3 mm, 3-4 mm, 1-3
Fig. 1 -Posterior scintignam of composite 6 mm, 50% background sub-
mm, and 1 -4 mm. Although the gamma camera-digital computer tracted 99”Tc DTPA nenognam (A). Kidneys were outlined via light pen (B)
system was started at the moment of injection, the subsequent and then either semilunar (C) on ring-shaped (D) background area was
sequential listing of calculated 1 5 sec net renal counts was used to created.
determine the arrival time of radiotracer within the kidneys. Tracer
arrival within the kidneys usually did not occur during the first 15 These variables of depth connection vs. no depth correction,
sec due to the central cardiopulmonary circulation of 99mTc-DTPA semilunar vs. ring background area, and multiple time intervals
which could be prolonged in individuals with cardiopulmonary ab- were used in all combinations in order to plot total percentage of
normalities. However, the arrival of radionuclide within the kidneys renal uptake of DTPA vs. 24 hr cneatinine clearance. A linear
was easily detected by inspecting the display of net renal counts. regression analysis was performed on a PDP-i 1 computer [4] in
The total net renal counts at the various time intervals were then order to determine the best combination of variables to be used for
plotted as a percentage of the administered counts vs. 24 hr rapid GFR calculation.
creatinine clearance: GFR determinations for each kidney were then computed once
% total renal uptake of DTPA =
the proper formula for calculating total GFA had been ascertained:
total counts = right counts + left counts; % right uptake = (right
(right kidney counts - background)
counts/total counts) x 1 00; and % left uptake = (left counts/total
+ (left kidney counts - background)
x 1 00. counts) X 1 00. Therefore, right GFR = (% right uptake) (total GFR)
preinjection counts - postinjection counts
and left GFR = (% left uptake) (total GFR).
A second correlation of percentage renal activity vs. 24 hr The fractional distribution of total GFR between the two kidneys
creatinine clearance was also determined using depth correction to at the subsequently computed proper time interval was plotted
compensate for gamma nay attenuation by the soft tissues inter- against the similar fractional distribution determined at 1 -3 mm,
posed between kidney center and scintillation camera. The formula which has been reported as an appropriate time for making such a
of Tonnesen et al. [2] was used for determining renal depth in differential determination [5]. A linear regression analysis was also
centimeters for normally positioned kidneys: right kidney depth =
performed on these data.
1 3.3 (weight/height) + 0.7 and left kidney depth = i 3.2 (weight/ In order to determine reproducibility of the calculation of glomen-
height) + 0.7, with weight in kilograms and height in centimeters. ulan filtration rate using this nadionuclide technique, a group of
The depth of transplanted kidneys was determined by sonogra- patients was studied twice and the results compared. The glomen-
phy. Once kidney depth was calculated, the corrected renal counts ulan filtration values of individual kidneys were plotted against each
were individually determined by the following formula: depth-cor- other and a linear regression analysis performed in order to deter-
rected rena! counts = (renal count - background/e” 1, where i mine the correlation of the paired studies.
is 0.1 53 (linear attenuation coefficient of 99mTc in soft tissues) [3],
and x is the kidney depth in centimeters. Thus the resulting depth-
corrected formula was: Results
fright kidney counts - background
% total renal uptake of DTPA = I Preliminary analysis of the results revealed the proper
\
time for data analysis to be 1 -2 mm, 2-3 mm, or 1 -3 mm
fleft kidney counts - background after 99mTc-DTPA arrival in the kidneys. The other time
+1
\ e intervals (i.e., 0-1 mm, 3-4 mm, and 1-4 mm) were less
x 100.
preinjection counts - postinjection counts accu rate.
AJR:1 38, March 1982 ESTIMATING GLOMERULAR FILTRATION WITH 99mTc-DPTA 567

TABLE 1 : Results of Linear Regression Analysis of Variables 120

Used in ComputIng GFR

Background. Time variables correlation Coefficient Standard Error 100

(min) (r) (5x.y)

With depth correction: z S

Semilunar: 80
1-2 0.942597 9.50768
2-3 0.953018 8.62453 S.

1-3 0.950688 8.83056 60 S

Ring:
C, S
1-2 . 0.931889 10.3280
2-3 . . . . 0.943694 9.41 905 2 40 #{149} #{149}#{149}#{149}
I

1-3 . . , 0.940269 9.69273 4 SI

Without depth correction:

S S
Semilunar: -. 20
4 I
1-2 . . . 0.864201 14.3255 U #{149}
2-3 . . 0.8#{176}2258 12.8558
1-3 . . 0.882933 13.3675 OS

Ring: 0 20 40 60 80 100 120

1-2 0.850409 14.9797

2-3 0.878286 13.61 34
1-3 . . 0.867935 14.1413 ACTUAL GFR (ML/MIN)
. Based on 33 studies
Fig. 3.-Calculated GFR (using final formula in text) plotted against actual
GFR (i.e., cneatinine clearance). Formula (based on V = a + b (X) with a =

-6.19843 and b = 9.75621) had r > 0.95 with Sx.y = 8.37.

12

10 kidneys was the best period for data analysis.

Consequently
the optimal combination correc-
of variables was: (1 ) depth
z tion; (2) semilunar background area of interest; (3) analysis
S
-I performed during the 2-3 mm time period after tracer arrival
within the kidneys. This combination had the best correlation

6
coefficient (r 0.95)
> and the lowest standard error (Sx . y
= 8.62) of any set of independent variables tested. Two
patients were added later to the series but analyzed using
4 only the semilunar background area. Figure 2 shows the
final correlation of percentage of total renal uptake of DTPA
2 using the specified best combination of variables versus 24
hr creatinine clearance. The final correlation coefficient
based on 35 studies was r > 0.95, the standard error =
a
0 20 40 60 80 100 120 8.37, regression coefficient = 9.75621 and intercept =

-6.19843.
The formula for calculation of GFR was derived from the
CIFATININE CLEARANCE (ML/MIN) results of the regression analysis: GFR mI/mm = (% total
renal DTPA uptake) (regression coefficient) + (intercept).
Fig. 2.-Percentage of total renal DTPA uptake at 2-3 mm plotted against
24 hr creatinine clearance values: 35 data points shown. Substitution into that formula yielded:
fright counts - background
‘,.. e’S3t33+O7I
Table 1 shows the results of the linear regression analysis (left counts - background
+*, e_O1S31132iW/H)+ OTt
correlating the variables of time, background configuration,
GFR
and depth correction with 24 hr creatinine clearance values. =
preinjectson counts - postinjection counts
These results were based on analyses of the first 33 ex-
x 100 x 9.75621 - 6.19843.
aminations in the series and covered a creatinine clearance
range of 1 -1 1 3 mI/mm. The introduction of depth correc- with W as weight and H as height.
tion into the analysis was more important than either the This formula was then used to compute the GFR from
configuration of the background area of interest or choice each patient’s original data. The resulting estimation of GFR
between the indicated time intervals. The semilunar back- was plotted against the 24 hr creatinine clearance using a
ground area resulted in a slightly better correlation coeffi- line of identity for comparison purposes (fig. 3). One calcu-
cient and lower standard error than the ring configuration. lation in a patient with a GFR of 1 mI/mm resulted in a
The 2-3 mm inclusive interval after tracer arrival in the negative number (-3.6) which was plotted as zero.
568 GATES AJR:138, March 1982

PERCENT INDIVIDUAL kIDNEY 120

DTPA UPTAKE

100 z
100

I 8
V
z

0 4 60
o 50
I V

40
0 z #{149}1.t

z
I
4 20
.5

o so ioo 0 2 4 6 8 10 12 14

2-3MIN WITH DEPTH CORRECTION SERUM CREATININE (MG/DL)

Fig. 4.-Percentage of individual kidney uptake of 99’Tc DTPA at 1 -3 mm Fig. 6.-Serum creatinine values plotted against 24 hr creatinine clear-
(not depth connected) vs. 2-3 mm (depth connected). Each dot is 1 data point, ance determinations in 35 patients. Small elevations in serum creatinine may
except circle which represents 2 superimposed points. Correlation of this be associated with marked decreases in 24 hr creatinine clearance. This
relationship was excellent (r > 0.99). nonlinear relation shows how serum creatinine values are not as sensitive to
small changes in renal function as are creatinine clearance determinations,

100
Another important factor to be used in analyzing any new
test is its reproducibility. Figure 5 shows the results of
individual kidney GFR calculations performed twice on the
same patients. These paired studies were done 24 hr apart
except for one which was at 48 hr. Individual kidney GFR
calculations were plotted rather than the sum of both kid-
neys (except for patients who had only a solitary functional
kidney or a transplanted kidney) since this would be a more
rigorous test of reproducibility. Conceivably, small variations
in individual renal GFR calculations could occur in paired
studies that might compensate for each other resulting in
the same total renal GFR determination yet vary in fractional
;j distribution of relative function. This series of 1 5 kidneys
(GFR range, 4-83 mI/mm) showed an excellent correlation
(r > 0.99) between paired studies, thus demonstrating re-
GFR (ML/MIN)
producibility of the technique.
Fig. 5.-Radionuclide computed GFR determinations performed twice in
1 5 kidneys by injections of DTPA on different days. Correlation (r > 0.99) of Discussion
these completely separate determinations demonstrated excellent neproduc-
ibility of technique. The serum creatinine determination is an accepted means
for evaluating gross renal function, yet it is relatively insen-
sitive to early significant changes in GFR. Creatinine clear-
Individual renal GFR calculations depend on determina- ance is a more sensitive parameter of renal function, as
tion of fractional distribution of 99mTc-DTPA between the two shown in figure 6 which plots the serum creatinine values of
kidneys at a specified time interval. A 1 -3 mm period has the 35 patients in this series vs. their 24 hr creatinine
been reported [5] as being appropriate for making this clearance determinations. Patients with serum creatinine
calculation on the basis of individual kidney creatinine clear- values in the range of 1 -2 mg/dl had clearance values of
ance determinations. The technique reported here uses a 27-1 1 3 mI/mm.
depth corrected, 2-3 mm interval for this computation. The determination of GFR became possible through the
Consequently the fractional uptake of 99mTc-DTPA among introduction of the clearance concept (i.e. , the volume of
62 kidneys using a non-depth corrected 1 -3 mm value was blood cleared of a substance by urine excreted in 1 mm)
plotted against the depth corrected, 2-3 mm value (fig. 4). and was initially determined by use of inulin, urea, or cre-
A linear regression analysis showed the correlation coeffi- atinine. Although inulin clearance is an ideal method for
cient for this relation to be r > 0.99. determining GFR [6-8], the technique has limited clinical
AJR:138, March 1982 ESTIMATING GLOMERULAR FILTRATION WITH 99mTc-DPTA 569

application because of the requirement for constant intra- Schlegel et al. [1 9] estimated GFR by determining the
venous inulin infusion during the procedure. Determination uptake ratios of either 99mTc iron ascorbate or 99mTc-Sn-
of creatinine clearance is simpler and has been adopted as complexed DTPA to ‘31I-hippurate. Ideally, the ratio of GFR
a clinical means of determining GFR. Creatinine clearance to ERPF (effective renal plasma flow which is determined
correlates closely with inulin clearance [9] but is not as from 131I-hippurate clearance) is 0.2 and is expressed as the
exact as inulin clearance in estimating GFR, since a small filtration fraction. However, this ratio in their normal patients
quantity of creatinine is secreted by the renal tubules [7]. was 0.35 when using iron ascorbate. Thus, the 99mTc/1311
Estimation of GFR via creatinine clearance is superior to uptake ratio was multiplied by 0.57 (i.e. , 0.2/0.35) in order
urea clearance determinations [7]. However, the creatinine to correct for the inconsistency between expected and
clearance test does have limitations. Creatinine clearance observed results. When using Sn-DTPA, the a9mTc/1311 up-
determines the total glomerular filtration rate (GFR) and not take ratio in their normal patients was 0.43. This was simi-
individual renal function unless the kidneys have been Se- larly corrected by multiplying the uptake ratios by 0.47 (i.e.,
lectively catheterized. Furthermore, an accurate 24 hr urine 0.20/0.43). By use of this ‘correction’ method, ‘ individual ‘

collection is essential ton precise results. Consequently, kidney and total GFR were calculated. The two different
radionuclide techniques have been developed in order to correction values could result from somewhat differing bio-
determine GFR both as a total value as well as individually pathways of these two a9mTcIabeled agents. Iron ascorbate
for each kidney. Various radiotracers have been used, in- is not entirely cleared by glomerular filtration (as is Sn-
cluding 14C-inulin, 51Cr-EDTA (ethylenediaminetetraacetic complexed DTPA) since there is some tubular retention of
acid), 251-iothalamate, 169Yb-DTPA, and 99mTc-DTPA. this agent.
DTPA is a soluble organic molecule of less than 500 The technique I describe has several advantages. The
molecular weight which is stable at physiologic pH. When correlation coefficient between DTPA uptake and creatinine
stannous (Sn) ion is mixed with pertechnetate, the che- clearance is as high or higher than that reported by others
lating properties of DTPA result in technetium stannous using plasma clearance techniques and is based on a large
DTPA. This radiopharmaceutical satisfies criteria for mea- patient population. The reproducibility of the method is
surement of GFR, since any such agent must be: (1 ) com- excellent (r > 0.99) as shown by comparing the results of
pletely filtered at the glomerulus; (2) not synthesized, de- repeated studies on the same patients performed on differ-
stroyed, reabsorbed or secreted by the renal tubule; (3) ent days. The fractional distribution of GFR between the two
physiologically inert; and (4) not be bound to plasma protein kidneys at the selected time interval has an excellent con-
[1 0]. Klopper et al. [1 1 ] confirmed that 99mTc-DTPA satisfies relation (r > 0.99) compared to the technique of Pieretti et
these requirements, although they did show minimal protein al. [5], who had a correlation coefficient of 0.96 when
binding that was considered insignificant. comparing the fractional accumulation of 9emTcDTPA vs.
Several radionuclide preparations and techniques have split creatinine clearance. Furthermore, this present tech-
been successfully used in order to calculate GFR. 51Cr- nique requires only a single radionuclide preparation and
EDTA is a valid agent for determining GFR by the single does not depend on a ‘corrected’
‘ ‘ filtration fraction value
injection method because its plasma clearance is nearly in order to determine GFR. The final advantage of the
identical to that of inulin [1 2-1 4]. This radiopharmaceutical technique is its speed and patient acceptance. The various
is primarily used in Europe, but it is not commercially avail- methods of determining plasma clearance of radiotracer
able in this country. Several single injection methods have require that several blood samples be drawn over several
been reported using 99mTc-DTPA for determining GFR [10, hours. However, the technique reported here needs only a
1 1 , 1 5-1 7]. All of these techniques require that serial blood few minutes of patient time and does not require any blood
samples be drawn at intervals up to 4 hr after tracer admin- samples-two conditions happily received in clinical prac-
istration in order to perform the exponential analysis of the tice.
plasma disappearance curve of 99mTc-DTPA from which
GFR is ultimately calculated. Braren et al. [1 0] compared
the radionuclide computed GFR values with 24 hr creatinine REFERENCES
clearance determinations in 24 patients and showed a sig- 1. VanSlyke DD, Hiller A, Miller BF. Clearance, extraction per-
nificant correlation (r = 0.87). Ash and Gilday [1 5] per- centage and estimated filtration of sodium ferrocyanide in
formed a similar analysis in 26 patients and also showed a mammalian kidney: comparison with inulin, creatinine and
good correlation (r = 0.91 ). Balachandran et al. [1 7] ob- urea. Am J Physiol 1935;1 13:611-628
tamed a fair correlation (r = 0.833) between 24 hr creatinine 2. Tgnnesen KH, Munck 0, Hald T, Mogensen P, Wolf H. Influ-
clearance and 9”Tc-DTPA computed GFR, but had a better ence on the renogram of variation in skin to kidney distance
correlation (r = 0.91 2) between 99mTc-DTPA computed GFR and the clinical importance hereof. Presented at the Interna-
tional Symposium Radionuclides in Nephrology, Berlin, April
and inulin clearance.
1974 (cited by Schlegel JU, Hamway SA. Individual renal
The high correlation of plasma clearance of commercial
plasma flow determination in 2 minutes. J Urol 1 976; 1 1 6 : 282-
99mTc-DTPA preparations (including the one used here: r>
285)
0.96) with 5tCr-EDTA plasma clearance has been reported 3. Early PJ, Razzak MA, Sodee DB. In: Textbook of nuclear
[1 6, 1 81. However, some variability occurred between dif- medicine technology, 3d ed. St. Louis: Mosby, 1979: 1 15
ferent preparations, and it would be wise to stay with a 4. Williamson M. COSAP, conversationally oriented statistical
single manufacturer in order to optimize reproducibility. analysis package. Appleton: Lawrence University, 1973
570 GATES AJR:138, March 1982

5. Pieretti A, Gilday D, Jells A. Differential kidney scan in pediatric tion rate measurement in man by the single injection method
urology. Urology 1974:4:665-668 using 5tCr-EDTA. Clin Sci 1969;37:169-180
6. PiUs RF. Clearance and rate of glomerular filtration. In: Physi- 1 4. Brbchner-Mortensen J, Giese J, Rossing N. Renal inulin clear-
ology of the kidney and body fluids, 2d ed. Chicago: Year ance versus total plasma clearance of 51Cr-EDTA. ScandJ Clin
Book Medical, 1968:62-70 Lab Invest 1969;23:301 -305
7. Meschan I. Background physiology of the urinary tract for the 1 5. Ash JM, Gilday DL. Renal nuclear imaging and analysis in
radiologist. Radiol Clin North Am 1965;3: 13-28 pediatric patients. Urol Clin North Am 1980;7:201-214
8. Berliner RW. Outline of renal physiology. In: Strauss MB, Welt 1 6. Hilson AJW, Mistry AD, Maisey MN. 99mTc DTPA for the
LG, eds. Diseases of the kidney, 2d ed. Boston: Little, Brown, measurement ofglomerular filtration rate. BrJ Radio! 1 976;49:
1971:31-85 794-796
9. Brod J, Sorota JH. Renal clearance of endogenous ‘ ‘creati- 1 7. Balachandran S, Toguni AG, Petrusick TW, Abbott LC. Com-
nine” in man. J Clin Invest 1948;27:645-654 parative evaluation of quantitative glomerular filtration rate
1 0. Braren V, Versage PN, Touya JJ, Bnill AB, Goddard J, Rhamy measured by isotopic and nonisotopic methods. Clin NucI Med
AK. Radioisotopic determination of glomerular filtration rate. J 1981 4: 150-1 53
Urol I 979;1 21 :145-147 1 8. Carlsen JE, Mgller ML, Lund JO, Trap-Jensen J. Comparison
11. Klopper JF, Hauser W, Atkins HL, Eckelman WC, Powell A. of four commercial Tc99m(Sn)DTPA preparations used for the
Evaluation of 99mTc-DTPA for the measurement of glomerular measurement of glomerular filtration rate: concise communi-
filtration
rate. J NucI Med 1972;1 3: 1 07-1 10 cation. J NucI Med 1980;21 :126-129
1 2. HaIl JE, Guyton AC, Farr BM. A single-injection method for 1 9. Schlegel JU, Halikiopoulos HL, Prima A. Determination of
measuring glomerular filtration rate. Am J Physiol 1977;232: filtration fraction using the gamma scintillation camera. J Urol
F72-F76 1979;1 22:447-450
1 3. Chantler C, Garnett ES, Parsons V, Veall N. Glomerular filtra-