Aim: To develop a strategy for checking the reliability of slope intercept measurement of glomerular filtration rate (GFR).
Methods: Six blood samples, obtained bilaterally 20-240 min after the injection of Cr-ethylenediaminetetraacetic acid (EDTA) and iohexol into opposite arms, were assayed for indicator injected contralaterally. GFR6iohexol, the reference value, was measured for all six samples and GFR3EDTA (slope-intercept) for the last three samples (per 1.73 m). GFR was measured from the half-time [GFR/extracellular fluid volume (ECV)3EDTA], for the single samples at 2-4 h (Christensen-Groth; GFR1) and for creatinine [estimated GFR (eGFR)].
Results: In six of the 97 studies, the correlation coefficient (r) of the fit to the last three sample points was less than 0.99. In the remaining 91 studies, GFR3EDTA disagreed with GFR/ECVEDTA by more than 15% in 19 studies (group A) and by less than 15% in 72 studies (group B). GFR3EDTA disagreed with GFR6iohexol by 12.6% in group A but only by 6.3% in group B (P<0.001). No such discrimination was displayed by eGFR. Although GFR3EDTA was within 15%of GFR6iohexol in 14 group A studies, eGFR was within 15% of GFR3EDTA in only eight studies and disagreed by more than 15% in six studies. Conversely, in the five of the 19 studies in which GFR3EDTA disagreed with GFR6iohexol by more than 15%, the agreement between eGFR and GFR3EDTA was less than 15% in two studies and more than 15% in three studies. GFR3EDTA was within 15% of GFR/ECV3EDTA in all six studies in which r was less than 0.99, and GFR3EDTA disagreed with GFR6iohexol by less than 15%. Confidence in GFR3EDTA was not improved by GFR1 or eGFR.
Conclusion: Slope intercept GFR is reliable if within 15% of GFR/ECV3. Estimated GFR was ineffective as a second checkpoint. GFR1 did not help when the fit was poor.