Measurement of Glomerular Filtration Rate - Journal of Nuclear Medicine

2 downloads 0 Views 806KB Size Report
In 1971, ihuxe, Maher, and Taylor introduced a single sample method for ... V. Yester, L. Keith Lloyd, W New!on Tauxe, and Eva V. Dubovsky. Division ofNuclear ...
Measurement of Glomerular Filtration Rate: Single InjectionPlasma Clearance Method Without Urine Collection Charles D. Russell, Patricia G. Bischoff, Frances N. Kontzen, Katherine L. Rowe!!, Michae! V. Yester, L. Keith Lloyd, W New!on Tauxe, and Eva V. Dubovsky Division ofNuclear Medicine, University ofAlabama Medical School: University ofAlabama at Birmingha,n; and Birmingham VA Medical Center, Birmingham, Alabama

Glomerular flftratlon rate (GFR)can be calculated from the plasma clearance of any of severalradIo@rmaceutIcals that are excreted by glomerularfiltration. Simplified methodshavebeen proposedthat require only one or two plasmasamplesin lieu of a more completeclearancecurve. Weexaminedthe error Introducedby this simplification. Forty patientswere studied using a dual-Isotopetechnique employing @mTcJDTPA and r69YbIDTPA,obtaining eight plasmasamplesfor each clearancecurve at Intervalsfrom 10 to 240 mm after injection. Data were fit to several empirical or semlempiricalformulae

and also to a two-compartmentcomputer model that permitted GFRestimationfrom only one or two data points. The computer model gave good fit, but so did severalsimpler methods.The error that resuftsfrom replacingthe completeclearancecurve by a single 3-hr samplewas about 8 mI/mm(residuals.d.). By using two samples(at I and 3 hr), the error could be reduced to 4 mI/mm. Recommended one- and two-sample methods are

presented.

J NuclMed26:1243-1247,1985

lomerular filtration rate (GFR) can be calculated from the rate of clearance of tracer activity from the plasma following a sing!e i.v. injection of a suitable ra diopharmaceutical. As long as the radiopharmaceutical is excreted solely by glomerular filtration and is not bound to plasma protein or to any other component of blood or other tissue, the GFR can be calculated simply by divid ing the administered dose by the integral of the plasma time-activity curve. Approximate methods have been pro

posed in which the GFR is estimated from only one or two plasma samples rather than from a multisample time-ac tivity curve. These have all been based on chromium-Si ethylenediaminetetraacetic

acid ([51Cr]EDTA), except for

the work of Jacobsson, who used technetium-99m diethy

binding (2). That problem was not addressed by Jacobs

son. In the present study, we employ [@Tc1DTPA, with explicit measurement of and correction for protein bind ing, and with simultaneous use of ytterbium-i69 (‘69Th) DTPA as an extra check.

In 1971, ihuxe, Maher, and Taylor introduced a single sample method for estimating effective renal plasma flow (ERPF) from the plasma clearance of iodine-13 1 orthoiodohippurate (3). Subsequent investigators have taken the same approach to estimate GFR from the plasma

clearance of [51Cr]EDTAor similar agents. Early investi gators fitted their data to purely empirical curves (4,5), while later ones often based their calculations on a one compartment open linear mathematical model (1,6-8).

acid ([99mTc]DTpA) (1). There is The latter approach augments the experimental data by a problem with the use of [@mTc]DTPAfor quantitative certain physiologic assumptions. Provided these assump measurements because of a variable degree of protein tions are valid, such a model can more accurately repre lenetriaminepentaacetic

sent the biological system, particularly for extrapolation Received Feb. 25, 1985; revision accepted Aug. 2, 1985.

outside the range of available experimental data. Two

For reprintscontact:CharlesD. Russell,MD, PhD, Div. of compartments are known to represent the data better than NuclearMedicine,UniversityofAlabama Hospitals,619-19th one compartment, but the implicit two-compartment St. South,Birmingham,AL 35233. problem cannot be solved analytically. We have treated

Volume26 • Number11 • November1985

1243

this problem by digital methods for the case in which only one or two data points are given. The results of the true two-compartment model are compared with previous simpler methods. Plasma clearance curves were measured for 40 patients by drawing eight blood samples from a heparin lock be tween 10 to 240 mm after simultaneous injection of [‘69Yb]DTPAand [99mTc]DTPA The GFR calculated

from all eight samples was regarded as the reference value and compared with results obtained by methods using only one or two samples. The resulting errors of measure

ment were tabulated for each agent, and the one-sample and two-sample methods that gave the best fit are described in detail. MATERIALS

AND METhODS

Patients were drawn from a population that needed GFR measurement for various clinical problems, includ

ing a group with normal renal function and recent spinal cord injuries, so that a wide range of GFR values could be included. Patients with edema, which alters radiotracer distribution (9), were excluded from the study. A butterfly infusion set was placed in a peripheral vein. Five millicuries of [@mTc]DTPAwere injected and the syringe flushed with blood, followed by 50 @tCiof [‘69YbJDTPA after which the syringe

was again flushed.

partment with proportionality constant a. V1 represented the volume of Compartment 1 and V2the volume of Corn partment 2. The one-point method employed fixed values for parameters a, V1, and V2, obtaining the GFR estimate for each patient by numerical fit (Newton's method) to the single data point. The fixed parameters a, V1, and V2, were chosen to give the best least squares fit for all pa tients taken together as a group. The two-point GFR was obtained in a similar fashion, using a/Vs and al V2 as fixed parameters, V1 and GFR as variable param eters, and Newton's method for the two-point, two-parameter curve fit.

RESULTS The result of using two data points to estimate GFR is shown in Fig. 1. The GFR calculated from two measure ments (at 1 and 4 hr) agreed closely with that calculated from all eight points (from iO mm to 4 hr after injection). A one-parameter formula initially proposed by Morgan et al. (5) was found to be equally satisfactory (Table 1). The two-parameter formula of Brøchner-Mortensen (11) also worked well when the parameters were chosen to fit our E'69YbIDTPA or [99mTc]DTPA data, though the pararn

eters reported by Brøchner-Mortensenfor a different. agent, [51Cr]EDTA,did not fit our data. A single sample drawn at 3 hr furnished a reasonable estimate of GFR that is probably adequate for most clini cal uses, though the error was about twice that of two

Residual activity in the syringe was less than 2 % of the dose. Standards were prepared by dilution from duplicate syringes. Eight blood samples were drawn into standard samples (Thble 2). The two-compartment EDTA-anticoagulatedvacuum sample tubes at 10, 20, 30, 45, 60, 120, 180, and 240 mm after injection, using a vein other than that used for injection. After centrifugation, duplicate samples of plasma and standard were pipetted, 0 counted, and the results averaged. A week later, after (l).E decay of @Tc,the samples were recounted for ‘69Yb. The aqueous standard solution of [‘69Yb]DTPA was pipet a.— ted into counting tubes within 8 hr of preparation, since

model again

z

further delay led to deposition of activity onto the glass