BACKGROUND: Measurements of GFR may be performed with a slope/intercept method (S/I), using only two blood samples taken in strictly defined time points. The aim of the study was to modify this method in order to extend time intervals suitable for blood sampling. Modification was based on a variation of a Russel et al. model parameter, selection of time intervals suitable for blood sampling and assessment of uncertainty of calculated results. MATERIAL AND METHODS: Archived values of GFR

more »

... ents of 169 patients with different renal function, from 5.5 to 179 mL/min, calculated with a multiple blood sample method were used. Concentrations of a radiopharmaceutical in consecutive minutes, from 60th to 190th after injection, were calculated theoretically, using archived parameters of biexponential functions describing a decrease in 99mTc-DTPA concentration in blood plasma with time. These values, together with injected activities, were treated as measurements and used for S/I clearance calculations. Next, values of S/I clearance were compared with the multiple blood sample method in order to calculate suitable values of exponent present in a Russel's model, for every combination of two blood sampling time points. A model was considered accurately fitted to measured values when SEE ≤ 3.6 mL/min. Assessments of uncertainty of obtained results were based on law of error superposition, taking into account mean square prediction error and also errors introduced by pipetting, time measurement and stochastic radioactive decay. RESULTS: The accepted criteria resulted in extension of time intervals suitable for blood sampling to: between 60 and 90 minutes after injection for the first sample and between 150 and 180 minutes for the second sample. Uncertainty of results was assessed as between 4 mL/min for GFR = 5-10 mL/min and 8 mL/min for GFR = 180 mL/min. CONCLUSIONS: Time intervals accepted for blood sampling fully satisfy nuclear medicine staff and ensure proper determination of GFR. Uncertainty of results is entirely acceptable and for high GFR values even comparable with uncertainty of multi-sample measurements.