Renal · GFR

Radionuclide GFR Measurement

Snapshot

Radionuclide GFR uses a filtered tracer (⁹⁹ᵐTc-DTPA or ⁵¹Cr-EDTA) with timed blood samples to give an accurate, reproducible glomerular filtration rate — more precise than creatinine-based estimates. It is used where accuracy matters, notably before and during nephrotoxic chemotherapy and before kidney donation.

A known amount of a purely filtered tracer is injected; its plasma clearance (from timed samples) equals GFR. This avoids the biases of creatinine-based equations.

Filtered tracerTrue clearance

Accurate> eGFR

Chemo/donationKey uses

Two plots of plasma activity after a filtered tracer: on a linear scale, the early fast component that the slope-intercept line misses is shaded; on a log scale, samples at 2, 3 and 4 hours define a straight line whose slope and back-extrapolated intercept give clearance. — Figure. Slope-intercept GFR: late samples (here 2, 3 and 4 h) lie on the slow, renal-clearance exponential, and its slope k and back-extrapolated intercept C₀ give clearance as dose × k ÷ C₀. Because this ignores the early distribution phase (shaded), it overestimates GFR and is corrected with the Brøchner-Mortensen equation before normalising to 1.73 m² (BNMS GFR guidelines, Fleming 2004).

When to measure

Before and during nephrotoxic chemotherapy (e.g. carboplatin dosing).
Living kidney-donor assessment.
When accurate GFR is needed and creatinine-based estimates are unreliable.

How it is done

Inject a filtered tracer (⁹⁹ᵐTc-DTPA or ⁵¹Cr-EDTA).
Timed venous samples: slope-intercept (2–4 samples, ~2–5 h) with Brøchner-Mortensen correction, or a single sample timed to expected GFR (2, 3 or 4 h) in adults without oedema/ascites.
Normalise to 1.73 m² body surface area (use absolute mL/min for carboplatin dosing).

Pitfalls

Accurate injection and sample timing are essential (avoid extravasation).
Oedema/ascites (expanded extracellular fluid) invalidate single-sample methods — use multi-sample methods.
Follow a validated sampling protocol; in severe renal impairment extend sampling (up to 24 h).

Evidence & guidelines

BNMS GFR guidelines (Fleming et al., Nucl Med Commun 2004;25:759–769; updated 2018).
More accurate than estimated GFR for dosing decisions.

In depth

⁵¹Cr-EDTA (not available in the USA) and ⁹⁹ᵐTc-DTPA are the usual radionuclide GFR agents, both with low radiation doses; plasma-sampling clearance has replaced urinary collection methods, which are prone to collection and bladder-emptying errors.
Among plasma methods, full multi-sample characterisation of the clearance curve (area under the curve) is the reference, but it is impractical routinely; the true gold standard, constant-infusion urinary clearance, is a research technique.
The slope-intercept method uses 2–4 samples taken roughly 2–5 h after injection; because it ignores the early fast exponential it overestimates GFR, so a Brøchner-Mortensen-type correction is applied, and the result is separately normalised to 1.73 m² body surface area.
The 2018 BNMS guideline recommends the single-sample method (Fleming formula) for routine GFR measurement in adults without ascites or oedema, with the sample time chosen according to the expected GFR.
In severe renal impairment, delayed sampling (up to 24 h) is needed because measurement error rises steeply at low GFR, and single-sample methods are invalid with expanded extracellular fluid (ascites, effusions, oedema), which requires multi-sample methods.
Camera-based GFR (Gates method, no blood sampling) is convenient but less accurate than plasma-sampling methods, so it should not be relied on when an accurate absolute GFR is required.
Serial GFR results must be interpreted against biological and measurement variability, which increases sharply at low GFR (especially below about 25 mL/min/1.73 m²), so small changes may not be significant.

Sources: Taylor, J Nucl Med 2014 · 55:608-15 (PMID 24549283) · BNMS GFR guideline, Fleming et al. Nucl Med Commun 2004 (PMID 15266169) · Taylor, J Nucl Med 2014 (PMID 24549283) · Murray et al. J Nucl Med Technol 2013 · 41:67-75 (PMID 23658207) · BNMS GFR guideline (Fleming et al. 2004, PMID 15266169) · BNMS GFR guideline 2018, as summarised in McMeekin et al. EJNMMI Phys 2022

← BackRenal