Model-Based Refinement of Transcutaneous FITC-Sinistrin GFR Estimation in Rats

Introduction: Transcutaneous devices now allow convenient GFR assessment in rats and mice, but concerns remain about their accuracy. This study aimed to improve GFR estimation by evaluating and addressing the limitations of current methods, especially errors arising from growth-related changes in body composition. Methods: Inbred Dahl salt-sensitive (SS) rats (male & female; 8 wk age, fed 0.4% salt diet (LS)) were studied. Under isoflurane anesthesia, two femoral venous catheters were placed in the inferior vena cava, and a carotid artery radiotelemetry catheter was implanted for continuous blood pressure monitoring. After a 1-week recovery period, rats were briefly re-anesthetized with isoflurane, and a transcutaneous GFR monitoring device (MediBeacon, Inc.) was positioned on the lateral abdominal skin. Once the animals had rapidly recovered from anesthesia, FITC-sinistrin (MediBeacon, Inc.) was administered through one of the venous catheters, followed by a 0.5 mL saline flush. Transcutaneous FITC fluorescence was then recorded at 1-second intervals, and approximately 50 μL of blood was collected from the second venous catheter at 2, 5, 10, 20, 40, and 90 minutes after FITC-sinistrin injection. The half-life (HL) of the FITC-sinistrin signal derived from the transcutaneous recording (TCT1/2, min) was calculated using the manufacturer’s software (MediBeacon). To determine the serum half-life (SeT1/2, min), blood samples were centrifuged, serum FITC fluorescence was quantified using a fluorescence plate reader (SPECTRA Fluor Plus), and SeT1/2 was calculated and compared with TCT1/2. The volume of distribution of sinistrin (Vd) was estimated from serum sinistrin concentrations measured at 20 and 40 minutes. Results: TCT1/2 and SeT1/2 agreed closely, with a mean bias of 0.54% and 95% limits of agreement −16.96% to 18.03%, indicating practical interchangeability. In 205 measurements, GFR × SeT1/2 was best fit by an age–body weight (BW) model: GFR × SeT1/2 = exp(5.36 − 0.0066·age + 0.84·ln(BW)), where age is in weeks and BW in kg (R 2 = 0.52). Adding sex, salt intake, or strain did not improve fit or Akaike's Corrected Information Criterion (AICc). Although salt intake did not improve the GFR × SeT1/2 model, it significantly (p< 0.05) affected Vd immediately after salt loading, and higher Vd was observed in the high salt group at a same BW. Log–log regression of GFR on BW yielded a common allometric slope (β ≈ 0.77) across sex groups, with differences limited to intercepts, consistent with vertical shifts in GFR rather than altered scaling. Conclusion: By applying this model, GFR can be predicted more precisely than with the current model. The assumption that GFR is proportional to body weight might cause over normalization. Small salt-induced changes did not modify the GFR x HL prediction model, possibly due to unequal group sizes, and GFR may need to be inferred indirectly from the Vd model. This abstract was presented at the American Physiology Summit 2026 and is only available in HTML format. There is no downloadable file or PDF version. The Physiology editorial board was not involved in the peer review process.