Renal Scanning 99mTc Diethylene-Triamine Pentaacetic Acid Glomerular Filtration Rate (GFR) Determination Compared With Iothalamate Clearance GFR in Diabetics

Renal Scanning 99mTc Diethylene-Triamine Pentaacetic Acid Glomerular Filtration Rate (GFR) Determination Compared With Iothalamate Clearance GFR in Diabetics Roger A. Rodby, MD, Amjad Ali, MD, Richard D. Rohde, Edmund J. Lewis, MD, and the Collaborative Study Group for The Study of Angiotensin-Converting Enzyme Inhibition in Diabetic Nephropathy • Traditionally, creatinine clearance is used as an estimation of the glomerular filtration rate (GFR) because of its relative ease and low cost. Errors in collection limit its usefulness. Estimation of GFR using 89mTc diethylene-triamine pentaacetic acid (Tc-DTPA) by direct scintigraphic determination of fractional radionuclide accumulation within each kidney does not require blood or urine sampling, takes 10 to 15 minutes to perform, and has been reported to give a GFR that correlates with 24-hour urinary creatinine clearance (CC) in hospitalized patients (r = 0.95). To assess its usefulness in the outpatient diabetic with nephropathy, 24 patients with type I diabetes underwent 56 iothalamate clearances during water diuresis and 56 simultaneous Tc-DTPA GFR estimations. GFR was also estimated from 24hour urinary CC, 100/creatinine, and by the formula of Cockcroft and Gault. Tc-DTPA GFR estimation by direct renal scanning correlated relatively poorly with iothalamate GFR (r = 0.74) in this patient population when all levels of iothalamate GFR were compared (n = 56), but improved (r = 0.80) when iothalamate GFR values greater than or equal to 120 mL/min were excluded from analysis (n = 45). Given all levels of iothalamate GFR, the best correlation was obtained with the estimation using the equation of Cockcroft and Gault (r = 0.86). . © 1992 by the National Kidney Foundation, Inc. INDEX WORDS: Glomerular filtration rate; creatinine clearance; Tc-DTPA; diabetic nephropathy; Cockcroft and Gault.

M

EASUREMENT of the glomerular filtration rate (GFR) in the diabetic patient is fraught with difficulties. Use of the serum creatinine overestimates the GFR by 10% to 15% because of creatinine secretion. In addition, the serum creatinine is affected by a number of extrarenal factors, including muscle mass and dietary intake of meat protein. 1 The formula of Cockcroft and Gault was developed to take into account the age-related decline in muscle mass, as well as gender- and weight-related differences in muscle mass. 2 The percent of total body mass that is represented by muscle in diabetics is often less than in nondiabetics, leading to the potential of an even greater overestimation of GFR when using the serum creatinine. Also, as GFR is lost, creatinine secretion increases and, therefore, the creatinine level may not increase until renal function is moderately reduced. 3 Finally, many patients with diabetes demonstrate glomerular hyperfiltration. These high GFRs are associated with low serum creatinine levels that are within the limits of sensitivity of the measurement of serum creatinine. A creatinine value of 0.7 mg/ dL with a ±0.1 mg/dL variance results in a range of estimated creatinine clearances (ECC) of 125 to 166 mL/min (using ECC = 100/serum creatinine) and 128 to 170 mL/min (using ECC by the formula of Cockcroft and Gault, 35-year-old male, weight 70 kg).

Twenty-four-hour urine collections for measurement of creatinine clearance if performed properly will correct for differing muscle masses, but are still affected by creatinine secretion. Diabetic autonomic neuropathy may affect the ability to empty the bladder properly and may limit this technique as well. Attempts to overcome this problem use water loading to increase urinary flow rates. This is usually performed in conjunction with the injection of a filtered, nonreabsorbed and nonsecreted radiolabeled isotope (eg, 125I_sodium lothalamate, 51Cr ethylene diamine tetraacetic acid [EDT A], 99mTc diethylene triamine pentaacetic acid [Tc-DTPAD for urinary clearance measurement as a gold standard of true GFR, and have been shown to correlate well with inulin clearance. 4,5 Plasma clearance of these radioisotopes has been used by many investigators to avoid the problems related with bladder empFrom the Departments of Medicine and Radiology, Rush Presbyterian-St. Luke's Medical Center, Chicago, IL. Received April 23, 1992; accepted in revised form August 14, 1992. Supported in part by US Public Health Service Grant No. 5-R01 DK39908, and in part by a grant from the BristolMeyers Squibb Institutefor Medical Research. Address reprint requests to Roger A. Rodby, MD, Section ofNephrology, Rush Presbyterian-St. Luke's Medical Center, 1653 W Congress Pkwy, Chicago, IL 60612. © 1992 by the National Kidney Foundation, Inc. 0272-6386/92/2006-0006$3.00/0

American Journal of Kidney Diseases, Vol XX, No 6 (December), 1992: pp 569-573

569

RODBY ET AL

570

tying and has been shown to correlate well with urinary clearance methods. 4 Although these plasma and urinary clearance methods give reasonably accurate estimations of GFR, they are cumbersome, take several hours to perform, require multiple measurements, and are usually limited to the research setting. Estimation ofGFR using Tc-DTPA by direct scintigraphic determination of fractional radionuclide accumulation within each kidney, occurring at a specific time interval after radionuclide administration, does not require blood or urine sampling, takes 10 to 15 minutes to perform, and has been reported to give GFR values that correlate with 24-hour urinary creatinine clearances (CC) in hospitalized patients with CC values that range from 5 to 110 mL/min (r = 0.95).6 To assess its usefulness in the outpatient diabetic with nephropathy, 24 patients with type I diabetes underwent 56 urinary iothalamate clearances during water diuresis and 56 simultaneous Tc-DTPA renal scanning GFR estimations. Renal function was also estimated from 24-hour urinary CC, 100/creatinine, and by the formula of Cockcroft and Gault.

METHODS Patients enrolled in the multicenter, double-blind, controlled trial of angiotensin-converting enzyme inhibition (ACEi) in type 1 diabetic nephropathy that participated through the Rush Presbyterian-St. Luke's Medical Center were used for this evaluation between February 1988 and November 1991. Patient inclusion criteria for entry into the ACEi study included the onset of diabetes mellitus before the age of 30 years, insulin dependence for a minimum of 7 years, age between 18 and 49 years, the presence of diabetic retinopathy, proteinuria of 500 mg/d or greater, and a serum creatinine concentration less than 221 ILmol/L (2.5 mg/dL). Serum and urinary creatinine measurements were performed using a Beckman Creatinine Analyzer II (Beckman Instruments, Fullerton CAl. Twenty-four-hour urine collections for CC were Table 1. Subject Data at Study Entry N

Male Female Age (yr) Weight (kg) Serum creatinine ILmol/L mg/dL lothalamate GFR (mL/min)

Mean ± SD

Range

11 13

..

,/

/

..

. ..... ,":

/

~;;

-'

;

.~.

lothalam&te CPR: (mllmlD)

lothalamate CPR: (mllmln)

lothllluraate CPR: (mVmln)

lothalamate CFR: (mVmJn)

Fig 1. Relationships of the different methods of measuring renal function compared with iothalamate GFR using linear regression. Mean regression lines, -; lines of identity, -----. obtained the day before the iothalamate clearance was performed . The serum creatinine level was used to estimate the CC by two methods: IOO/creatinine, and by the equation of Cockcroft and Gault: (140 - age [yr] X (weight [kgJ)/(72) X (serum creatinine [mgjdLJ), X .85 for women. Four 1251_ iothalamate urinary clearance periods using the single injection technique (without epinephrine) during water diuresis were performed. 7 The desired urinary flow rate was 3 to 10 mL/ min. Clearances were performed in the morning, after a normal breakfast, and were done with patients in the sitting position. Clearance periods averaged 30 minutes in duration. The interperiod coefficients of variation averaged 17%. Serum and urine samples obtained during the water diuresis were also analyzed for creatinine to provide a simultaneous CC measurement. Immediately following the end of the water diuresis iothalamate/CC study, Tc-DTPA GFR measurements were performed after the intravenous injection of Tc-DTPA by the renal scintigraphic technique as described by Gates. 6 Total GFR was reported as the sum of the right and left kidney GFR measurements. The Tc-DTPA measurement was compared with the iothalamate GFR and the simultaneous Cc. In addition, the 24-hour urinary CC, the ECC by 100/serum creatinine, the ECC by the equation of Cockcroft and Gault, and the simultaneous CC were similarly compared with the iothalamate GFR. Comparisons were made using linear and polynomial regression models with correlation coefficients (r values) and the standard deviations around the regression line (SDR) as determined by the method of least squares.

RESULTS 34 ± 6.8 78 ± 15

22-51 55-106

111 ± 41 1.26 ± 4. 6

62-203 .7-2.3

87 ± 40

19-166

Patient characteristics and renal functional data at the time of inclusion into the study are summarized in Table 1. Urinary flow rates averaged 7.5 mL/min ± 3.5 SO (3 to 17 mL/min). Figure I compares the relationships of each measurement or estimation of renal function to

Tc-DTPA GFR DETERMINATION IN DIABETICS

571

iothalamate GFR using linear regression. As demonstrated, the poorest correlation exists with the GFR as measured by Tc-DTPA (r = 0.74). The best correlation exists with the estimated clearance by the equation of Cockcroft and Gault (r = 0.86). All methods tend to overestimate the GFR at lower levels of GFR «40 mL/min). In addition, Tc-DTPA GFR and 100/creatinine tend to underestimate GFR at higher levels of GFR (~120 mL/min). Inspection of the TcDTPA versus iothalamate GFR graph demonstrates that as GFR increases above 120 mL/min, Tc-DTPA GFR underestimates iothalamate GFR and there is a decline in the accuracy ofTcDTPA as a measure of GFR. Recalculation of the correlation coefficient excluding all iothalamate GFRs greater than or equal to 120 mL/min improves the correlation coefficient somewhat (r = 0.80), but is still inferior to all other methods ofGFR determination that were evaluated. Taking into account this underestimation ofGFR by Tc-DTPA at the higher GFR values, the two methods of GFR determination were compared using polynomial regression analysis. This is demonstrated in Fig 2. This method of comparison significantly improves the correlation compared with the linear regression model (r = 0.78 SDR = 18.8, P = 0.0052) when all levels ofGFR are compared, but does not improve it at GFR values less than or equal to 120 mL/min. Table 2 gives the correlation coefficients and the stan-

IMtl

C

~

E

1"0 140

120

0:1

100

..

,,"

m a,~'

U

80

00( Q.

60

Q

40

Eo-

.:.

Eo-

II

•a

•

•

y=I1+1.3x .. 0043x 2 r=.78

20

20

40

60

80

lothalamate

100

GFR:

120

140

160

Tc-DTPA 1DO/Creatinine 24-Hour urinary CC Cockroft and Gault

= 56) SDR

0.74 0.78' 0.78 0.84 0.86

20.0 18.8' 18.7 22.3 20.9

SDR

0.80

16.5

0.88 0.85 0.83

14.4 20.2 20.7

, As determined by polynomial regression (Fig 2); all other values obtained by linear regression (Fig. 1).

dard deviations around the regression line for each GFR comparison with all GFR data (n = 56) and with the exclusion of GFR greater than or equal to 120 mL/min (n = 45). The simultaneous CC obtained during water diuresis correlated extremely well (r = 0.96 SDR = 11.6) with the iothalamate GFR, but overestimated it by approximately 10 mL/min at most levels of renal function. Comparison of the 24hour urinary CC to the simultaneous CC did not correlate as well (r = 0.83 SDR = 23.8). Comparison of the Tc-DTPA GFR values to the simultaneous CC values did not improve the correlation coefficients (Tc-DTPA vall simultaneous CC: r = 0.74, SDR = 20.3; Tc-DTPA v simultaneous CC ::5: 120 mL/min: r = 0.75, SDR = 17.7). DISCUSSION

••

•

All GFRs (n lothalamate GFR versus:

GFRs" 120 mL/min (n = 45)

,

_B_-=-_~

-iii "

0:1

~

,

,,

a,"'"

.. C

.

,,

,,

Table 2. Correlation Coefficients and Standard Deviations Around the Regression Line With Different Ranges of GFR

IIW

(mJlmin)

Fig 2. Comparison of Tc-DTPA GFR to iothalamate GFR. using .poly~omial regression. Mean regression line, -; hne of Identity, ----.

It is not surprising that each of these methods of GFR estimation are overestimations when compared with the iothalamate GFR, since they are all based on the CC, which overestimates GFR because of creatinine secretion. In addition, although Tc-DTPA GFR by direct renal scanning is reported to have a very high correlation coefficient (r = 0.95),6 this is compared with CC and therefore would also be expected to overestimate GFR. Still, it had the poorest correlation coefficient of all the methods of GFR determination studied when compared with the iothalamate GFR. The correlation is especially poor when the GFR is greater than or equal to 120 mL/min. It is worth noting that our study included a number

RODBY ET AL

572

of diabetics with GFR values that are in the range that is considered hyperfiltration. Patients with these high GFR values were not included in the study published by Gates. 6 There are two possible explanations for the TcDTPA GFR underestimation of iothalamate GFR in the normal to supranormal GFR range. Either the Tc-DTPA GFR measurement is unreliable at these levels of GFR, or perhaps the iothalamate GFR measurements are overestimations at these higher levels ofGFR. Little data are available comparing iothalamate GFR determinations by the single injection method to inulin clearance values in patients with normal and supranormal GFR values. Perrone et al found that GFR determinations using iothalamate clearance appeared to overestimate GFR compared with inulin clearance values in this higher GFR range (n = 4),5 while Israelit et al found that it tended to underestimate GFR (n = 4).7 Although the reliability of this iothalamate clearance technique as a method of GFR determination in these higher GFR values is yet to be determined, there is no apparent reason to assume that its correlation to inulin clearance deteriorates at these higher levels of GFR. The relatively good correlation of the calculated CC as obtained by the equation of Cockcroft and Gault to the iothalamate GFR makes this determination useful, since it requires only a single serum creatinine determination. That it seems to correlate better than the 24-hour urinary CC probably reflects the difficulty in attaining a proper collection. In addition, 24-hour urinary CC values are subject to circadian, dietary, and postural influences that are less likely to affect the iothalamate clearance values. Lemann et al reported similar findings with CC as determined by Cockcroft and Gault as compared with iothalamate GFR in a population of both diabetics with renal disease and patients with other kidney disorders. 8 These findings were confirmed by DeSanto et al in a population of both normal and glomerulopathic subjects. 9 The usefulness of this correlation should not be underestimated, since it appears to give a reasonable estimate of GFR over all ranges of GFR. The GFR value obtained by this method may be adequate for many clinical decisions. Although estimations based on creatinine values tend to overestimate the GFR because of creatinine se-

cretion, the clinical relevance of this overestimation is unknown. The use of a blocker of creatinine secretion (eg, cimetidine) before creatinine measurements are made has been demonstrated to give values similar to that obtained with inulin clearance and may be useful where more exact GFR measurements are necessary.IO,11 Renal functional data in interventional trials are usually obtained through more accurate measurements of GFR that depend on urinary clearances of substances that are handled similar to inulin. Although cumbersome and expensive, this process will continue unless it is determined that indirect measurements of renal function, despite their inaccuracies, are reliable enough to adequately determine changes in renal function (along longitudinal patient data) that give conclusions similar to that when direct measurements of renal function are used. The potential advantage of the Tc-DTPA GFR by direct renal scanning relates to its ease: it takes only about 10 to 15 minutes to perform, and neither blood nor urine need be collected. This would be especially useful in the diabetic, where diabetic cystopathy may prevent accurately timed urine collections. Unfortunately, the data presented in this study have demonstrated that this method of GFR measurement has limited usefulness in the diabetic patient with evidence of nephropathy. Other methods of determination are better indicators of GFR over the entire range of renal function in the diabetic patient. APPENDIX Members of the Collaborative Study Group include Clinical Coordinating Center: Edmund J. Lewis, MD (Principal Investigator), Rush Presbyterian-St. Luke's Medical Center, Chicago, IL; Lawrence Hunsicker, MD (Co-Principal Investigator), University ofIowa, Iowa City, IA; Biostatistical Coordinating Center: Raymond Bain, PhD (Principal Investigator), John Lachin, ScD (Co-Principal Investigator), George Washington University, Bethesda, MD; Central Laboratory: Susan H. Hou, MD (Director, Clinical Pathology), Melvin M. Schwartz, MD (Director, Anatomic Pathology), Rush Presbyterian-St. Luke's Medical Center, Chicago, IL; Clinic Directors: J. B. Wish, MD, J. Sheehan, MD, Case Western Reserve University, Cleveland, OH; M. A. Pohl, MD, Cleveland Clinic, Cleveland, OH; T. Berl, MD, University of Colorado, Denver, CO; G. Santiago, MD, Henry Ford Hospital, Detroit, MI; L. Hunsicker MD, University ofIowa, Iowa City, IA; J. Lemann, Jr, MD, S. S.Blumenthal, MD, Medical College of Wisconsin, Milwaukee, WI; L. A. Hebert, MD, N. S. Nahman, Jr, MD, Ohio State University, Columbus, OH; S. Goldfarb, MD, S. Kobrin, MD, University of Pennsylvania,

573

Tc-DTPA GFR DETERMINATION IN DIABETICS Philadelphia, PA; R. Rodby, MD, Rush Presbyterian-St. Luke's Medical Center, Chicago, IL; A. Levey, MD, M. L. McLaughlin, MD, M. E. Williams, MD, New England Medical Center/Joslin Diabetes Center, Boston, MA; J. McGill, MD, J. V. Santiago, MD, Washington University, St. Louis, MO; F. C. Whittier, MD, Affiliated Hospitals of Canton, Y oungstown, OH; D. C. Cattran, MD, University of Toronto, Toronto, Ontario; S. Lietz, MD, University of Illinois, Chicago, IL; J. Hano, MD, Loyola Medical Center, Maywood, IL; D. Maxwell, MD, Indiana University, Indianapolis, IN; J. G. Porush, MD, S. Spitalewitz, MD, Brookdale Hospital Medical

Center, Brooklyn, NY; K. Shapiro, MD, Nyack Hospital, Nyack, NY; S. Adler, MD, Harbor-UCLA Medical Center, Torrance, CA; N. Tolchin, DO, Nephrology and Internal Medicine Specialists, PC, Syracuse, NY; W. E. Hoy, MB, BS, Lovelace Medical Foundation, Albuquerque, NM; L. P. Svetkey, MD, Duke University, Durham, NC; Z. Sharon, MD, B. J. Rosenbaum, MD, Atlanta Nephrology Referral Center, Decatur, GA; J. Anolick, MD, Larchmont Medical Center, Mount Laurel, NJ; H. D. Tildesley, MD, R. Rangno, MD, St. Pauls Hospital, Vancover, BC; C. Joyce, MD, Memorial University of Newfoundland.

REFERENCES 1. Levey AS, Perrone RD, Madias NE: Serum creatinine and renal function. Annu Rev Med 39:465-490, 1988 2. Cockcroft SW, Gault HM: Prediction of creatinine clearances from serum creatinine. Nephron 16:31-41, 1986 3. Shemesh 0, Golbetz H, Kriss JP, et al: Limitations of creatinine as a filtration marker in g1omerulopathic patients. Kidney Int 28:830-838, 1985 4. Levey AS: Measurement of renal function in chronic renal disease. Kidney Int 38:167-184, 1990 5. Perrone RD, Steinman TI, Beck GJ, et al: Utility of radioisotopic filtration markers in chronic renal insufficiency: Simultaneous comparison of '25I-iothalamate, '69Yb_DTPA, 99mTc_DTPA, and inulin. Am J Kidney Dis 16:224-235,1990 6. Gates GF: Glomerular filtration rate: Estimation from fractional renal accumulation of 99mTc_DTPA (stannous). Am J Radiol 138:565-570, 1982

7. Israelit AH, Long DL, White MG, et al: Measurement of glomerular filtration rate utilizing a single subcutaneous injection of 125I-iothalamate. Kidney Int 4:346-349, 1973 8. Lemann J, Bidani AK, Bain RP, et al: Use of the serum creatinine to estimate glomerular filtration rate in health and early diabetic nephropathy. Am J Kidney Dis 6:236-243, 1990 9. DeSanto NG, Coppola S, Anastasio P, et al~ Predicted creatinine clearance to assess glomerular filtration rate in chronic renal disease in humans. Am J Nephrol 11:181-185, 1991 10. Hilbrands LB, Artz MA, Wetzels JFM, et al: Cimetidine improves the reliability of creatinine as a marker of glomerular filtration. Kidney Int 40: 1171-1176, 1991 11. Roubenoff R, Drew H, Moyer M, et al: Oral cimetidine improves the accuracy and precision of creatinine clearance in lupus nephritis. Ann Intern Med 113:501-506, 1990