Estimation of glomerular filtration rate with 99Tc(m)-DTPA: a comparative assessment of simplified methods

In 40 adult patients undergoing gamma camera renography, glomerular filtration rate (GFR) was measured using simplified 99Tc(m)-DTPA methods (i.e. a personal modification of the 'slope' method which does not require dose calibration, Gates' method and Carlsen's method) and compared to reference results (obtained using Sapirstein's formula and Russell's two-sample method with 51Cr-EDTA). Estimation of GFR from plasma creatinine (the Cockroft-Gault formula) was also carried out. Bias and imprecision of the simplified estimates were determined by the Bland-Altman method. The GFR values of the 'slope' method correlated best with the reference values (R2 = 0.88, S.E.E. = 11.3 ml min[-1]). Correlation of the two methods based on external determination with the gamma camera was no better at estimating GFR than that from plasma creatinine. Moreover, Gates' method underestimated GFR at all levels between 25 and 150 ml min(-1), while Carlsen's method overestimated at low levels and underestimated at high levels. The bias was as follows (ml): Cockroft-Gault 2.4; 'slope' -4.1; Carlsen 7.5; Gates 16.7. The imprecision was as follows (ml): 'slope' 11.8; Cockroft-Gault 16.4; Carlsen 20.5; Gates 22.8. We conclude that our modification of the slope method correlated best with the reference results, and would appear suitable for routine practice because of the small error involved. When performing sequential renal scintigraphy, it can also be used for a quick check of dubious data based on gamma camera methods.